LLM Interpretability

LLM: Large Language Model

Follow this topic Ask about this topic

Large language model interpretability asks what information models represent, how it moves through layers and components, and which computations cause particular answers or behaviors. Researchers use probes and representation geometry, sparse autoencoders, attribution and circuit tracing, activation patching or ablation, and steering to study factual recall, mathematical reasoning, hallucination, refusal, and in-context learning. These methods can reveal distributed circuits and layer-wise signals, and sometimes support targeted interventions without retraining. A central challenge is that information being decodable—or a direction being steerable—does not prove the model uses it; causal faithfulness, robustness across prompts and model families, and reliable evaluation remain open tensions.

Latest papers

781 papers82 papers in the last four weeks, up 12% on the four weeks before. 0.5% of all new papers.

Apr 15, 2026

Applied Explainability for Large Language Models: A Comparative Study

Large language models (LLMs) achieve strong performance across many natural language processing tasks, yet their decision processes remain difficult to interpret. This lack of transparency creates challenges for trust, debugging, and deployment in real-world systems. This paper presents an applied comparative study of three explainability techniques: Integrated Gradients, Attention Rollout, and SHAP, on a fine-tuned DistilBERT model for SST-2 sentiment classification. Rather than proposing new methods, the focus is on evaluating the practical behavior of existing approaches under a consistent and reproducible setup. The results show that gradient-based attribution provides more stable and intuitive explanations, while attention-based methods are computationally efficient but less aligned with prediction-relevant features. Model-agnostic approaches offer flexibility but introduce higher computational cost and variability. This work highlights key trade-offs between explainability methods and emphasizes their role as diagnostic tools rather than definitive explanations. The findings provide practical insights for researchers and engineers working with transformer-based NLP systems. This is a preprint and has not undergone peer review.

Venkata Abhinandan Kancharla

NRI Institute of Technology

LLM Interpretability Explainability Evaluation Transformer Interpretability Feature Attribution Faithfulness of Language Model Explanations

Apr 14, 2026

Hessian-Enhanced Token Attribution (HETA): Interpreting Autoregressive LLMs

Attribution methods seek to explain language model predictions by quantifying the contribution of input tokens to generated outputs. However, most existing techniques are designed for encoder-based architectures and rely on linear approximations that fail to capture the causal and semantic complexities of autoregressive generation in decoder-only models. To address these limitations, we propose Hessian-Enhanced Token Attribution (HETA), a novel attribution framework tailored for decoder-only language models. HETA combines three complementary components: a semantic transition vector that captures token-to-token influence across layers, Hessian-based sensitivity scores that model second-order effects, and KL divergence to measure information loss when tokens are masked. This unified design produces context-aware, causally faithful, and semantically grounded attributions. Additionally, we introduce a curated benchmark dataset for systematically evaluating attribution quality in generative settings. Empirical evaluations across multiple models and datasets demonstrate that HETA consistently outperforms existing methods in attribution faithfulness and alignment with human annotations, establishing a new standard for interpretability in autoregressive language models.

Vishal Pramanik, Maisha Maliha, Nathaniel D. Bastian +1

Department of Computer & Information Science & Engineering, University of Florida, Gainesville, FL 32611, USA · School of Computer Science, University of Oklahoma, Norman, OK 73019, USA · Department of Electrical Engineering and Computer Science, United States Military Academy, West Point, NY 10996, USA

Feature Attribution LLM Interpretability Faithfulness of Language Model Explanations Decoder-Only Language Models

Apr 14, 2026

IDEA: An Interpretable and Editable Decision-Making Framework for LLMs via Verbal-to-Numeric Calibration

Large Language Models are increasingly deployed for decision-making, yet their adoption in high-stakes domains remains limited by miscalibrated probabilities, unfaithful explanations, and inability to incorporate expert knowledge precisely. We propose IDEA, a framework that extracts LLM decision knowledge into an interpretable parametric model over semantically meaningful factors. Through joint learning of verbal-to-numerical mappings and decision parameters via EM, correlated sampling that preserves factor dependencies, and direct parameter editing with mathematical guarantees, IDEA produces calibrated probabilities while enabling quantitative human-AI collaboration. Experiments across five datasets show IDEA with Qwen-3-32B (78.6%) outperforms DeepSeek R1 (68.1%) and GPT-5.2 (77.9%), achieving perfect factor exclusion and exact calibration -- precision unattainable through prompting alone. The implementation is publicly available at https://github.com/leonbig/IDEA.

Yanji He, Yuxin Jiang, Yiwen Wu +3

The Hong Kong University of Science and Technology (Guangzhou) · Huawei Technologies Co.,Ltd · The Hong Kong University of Science and Technology

LLM Decision-Making Language Model Calibration LLM Interpretability Human-AI Decision Making

Apr 14, 2026

Latent Planning Emerges with Scale

LLMs can perform seemingly planning-intensive tasks, like writing coherent stories or functioning code, without explicitly verbalizing a plan; however, the extent to which they implicitly plan is unknown. In this paper, we define latent planning as occurring when LLMs possess internal planning representations that (1) cause the generation of a specific future token or concept, and (2) shape preceding context to license said future token or concept. We study the Qwen-3 family (0.6B-14B) on simple planning tasks, finding that latent planning ability increases with scale. Models that plan possess features that represent a planned-for word like "accountant", and cause them to output "an" rather than "a"; moreover, even the less-successful Qwen-3 4B-8B have nascent planning mechanisms. On the more complex task of completing rhyming couplets, we find that models often identify a rhyme ahead of time, but even large models seldom plan far ahead. However, we can elicit some planning that increases with scale when steering models towards planned words in prose. In sum, we offer a framework for measuring planning and mechanistic evidence of how models' planning abilities grow with scale.

Michael Hanna, Emmanuel Ameisen

ILLC, University of Amsterdam · Anthropic

LLM Planning Implicit Reasoning in Language Models LLM Interpretability Language Model Scaling

Apr 13, 2026

Phantom Paradoxes: Language Models Treat Unresolved Circular Truth Dependence as Contradiction

Do language models struggle with self-reference itself, or with structures where propagating truth values along references cannot settle a value in finitely many steps? Our results support the latter. In our main test, statements refer to one another's truth in a cycle, yet exactly one consistent assignment exists. Local propagation never settles the queried value, but exhaustive search finds it, so INCONSISTENT (no consistent assignment) is always wrong. We call such misreadings of satisfiable problems phantom paradoxes. Each circular problem has a grounded twin with the same answer but no cycle. To rule out logical difficulty and unfamiliar wording, we added constraint twins (equivalent formulas without talk of truth) and perplexity-matched grounded controls. In confirmatory behavioral tests fixed before any confirmation response, four open models from three families answered INCONSISTENT on average 31.0 percentage points more often on circular problems than on grounded twins, 36.4 more than on constraint twins and 37.1 more than on perplexity-matched controls. In each model, validated layer-wise truth directions, fitted on unrelated arithmetic, recovered the queried value at the question's last token for grounded twins but not for circular problems. Constraint twins, likewise unrecovered, drew INCONSISTENT far less often, so circular truth dependence, not the unrecovered value alone, matters. In a small natural-language bank, self-reference settling in finitely many steps drew INCONSISTENT 23.4 points less often than unsettled self-reference with the same classical solution. Pooled tests did not show the predicted late-layer revision or attention spread. Overall, models treat unresolved circular truth dependence as contradiction. We will release all prompts, outputs and hidden states.

Ji Ho Bae

JRTI

LLM Interpretability LLM Reliability

Apr 13, 2026

Robust Explanations for User Trust in Enterprise NLP Systems

Robust explanations are increasingly required for user trust in enterprise NLP, yet pre-deployment validation is difficult in the common case of black-box deployment (API-only access) where representation-based explainers are infeasible and existing studies provide limited guidance on whether explanations remain stable under real user noise, especially when organizations migrate from encoder classifiers to decoder LLMs. To close this gap, we propose a unified black-box robustness evaluation framework for token-level explanations based on leave-one-out occlusion, and operationalize explanation robustness with top-token flip rate under realistic perturbations (swap, deletion, shuffling, and back-translation) at multiple severity levels. Using this protocol, we conduct a systematic cross-architecture comparison across three benchmark datasets and six models spanning encoder and decoder families (BERT, RoBERTa, Qwen 7B/14B, Llama 8B/70B; 64,800 cases). We find that decoder LLMs produce substantially more stable explanations than encoder baselines (73% lower flip rates on average), and that stability improves with model scale (44% gain from 7B to 70B). Finally, we relate robustness improvements to inference cost, yielding a practical cost-robustness tradeoff curve that supports model and explanation selection prior to deployment in compliance-sensitive applications.

Guilin Zhang, Kai Zhao, Jeffrey Friedman +3

Workday AI Research

Explainable Artificial Intelligence LLM Interpretability Explanation Stability Explainability Evaluation

Apr 13, 2026

Psychological Concept Neurons: Can Neural Control Bias Probing and Shift Generation in LLMs?

Using psychological constructs such as the Big Five, large language models (LLMs) can imitate specific personality profiles and predict a user's personality. While LLMs can exhibit behaviors consistent with these constructs, it remains unclear where and how they are represented inside the model and how they relate to behavioral outputs. To address this gap, we focus on questionnaire-operationalized Big Five concepts, analyze the formation and localization of their internal representations, and use interventions to examine how these representations relate to behavioral outputs. In our experiment, we first use probing to examine where Big Five information emerges across model depth. We then identify neurons that respond selectively to each Big Five concept and test whether enhancing or suppressing their activations can bias latent representations and label generation in intended directions. We find that Big Five information becomes rapidly decodable in early layers and remains detectable through the final layers, while concept-selective neurons are most prevalent in mid layers and exhibit limited overlap across domains. Interventions on these neurons consistently shift probe readouts toward targeted concepts, with targeted success rates exceeding 0.8 for some concepts, indicating that the model's internal separation of Big Five personality traits can be causally steered. At the label-generation level, the same interventions often bias generated label distributions in the intended directions, but the effects are weaker, more concept-dependent, and often accompanied by cross-trait spillover, indicating that comparable control over generated labels is difficult even with interventions on a large fraction of concept-selective neurons. Overall, our findings reveal a gap between representational control and behavioral control in LLMs.

Yuto Harada, Hiro Taiyo Hamada

Araya Inc.

Personality Modeling in Language Models LLM Interpretability Causal Interventions in Language Models

Apr 13, 2026

A Mechanistic Analysis of Looped Reasoning Language Models

Reasoning has become a central capability in large language models. Recent research has shown that reasoning performance can be improved by looping an LLM's layers in the latent dimension, resulting in looped reasoning language models. Despite promising results, few works have investigated how their internal dynamics differ from those of standard feedforward models. In this paper, we conduct a mechanistic analysis of the latent states in looped language models, focusing in particular on how the stages of inference observed in feedforward models compare to those observed in looped ones. To this end, we analyze cyclic recurrence and show that for many of the studied models each layer in the cycle converges to a distinct fixed point; consequently, the recurrent block follows a consistent cyclic trajectory in the latent space. We provide evidence that as these fixed points are reached, attention-head behavior stabilizes, leading to constant behavior across recurrences. Empirically, we discover that recurrent blocks learn stages of inference that closely mirror those of feedforward models, repeating these stages in depth with each iteration. We study how recurrent block size, input injection, and normalization influence the emergence and stability of these cyclic fixed points. We believe these findings help translate mechanistic insights into practical guidance for architectural design.

Hugh Blayney, Álvaro Arroyo, Johan Obando-Ceron +4

University of Oxford · Mila – Quebec AI Institute · Universit´e de Montréal +1

Recurrent Transformers LLM Interpretability Implicit Reasoning in Language Models Recurrent Neural Networks

Apr 13, 2026

Why Do Large Language Models Generate Harmful Content?

Large Language Models (LLMs) have been shown to generate harmful content. However, the underlying causes of such behavior remain under explored. We propose a causal mediation analysis-based approach to identify the causal factors responsible for harmful generation. Our method performs a multi-granular analysis across model layers, modules (MLP and attention blocks), and individual neurons. Extensive experiments on state-of-the-art LLMs indicate that harmful generation arises in the later layers of the model, results primarily from failures in MLP blocks rather than attention blocks, and is associated with neurons that act as a gating mechanism for harmful generation. The results indicate that the early layers in the model are used for a contextual understanding of harmfulness in a prompt, which is then propagated through the model, to generate harmfulness in the late layers, as well as a signal indicating harmfulness through MLP blocks. This is then further propagated to the last layer of the model, specifically to a sparse set of neurons, which receives the signal and determines the generation of harmful content accordingly.

Rajesh Ganguli, Raha Moraffah

Worcester Polytechnic Institute

LLM Safety LLM Interpretability

Apr 13, 2026

METER: Evaluating Multi-Level Contextual Causal Reasoning in Large Language Models

Contextual causal reasoning is a critical yet challenging capability for Large Language Models (LLMs). Existing benchmarks, however, often evaluate this skill in fragmented settings, failing to ensure context consistency or cover the full causal hierarchy. To address this, we pioneer METER to systematically benchmark LLMs across all three levels of the causal ladder under a unified context setting. Our extensive evaluation of various LLMs reveals a significant decline in proficiency as tasks ascend the causal hierarchy. To diagnose this degradation, we conduct a deep mechanistic analysis via both error pattern identification and internal information flow tracing. Our analysis reveals two primary failure modes: (1) LLMs are susceptible to distraction by causally irrelevant but factually correct information at lower level of causality; and (2) as tasks ascend the causal hierarchy, faithfulness to the provided context degrades, leading to a reduced performance. We belive our work advances our understanding of the mechanisms behind LLM contextual causal reasoning and establishes a critical foundation for future research. Our code and dataset are available at https://github.com/SCUNLP/METER .

Pengfeng Li, Chen Huang, Chaoqun Hao +4

College of Computer Science, Sichuan University · Engineering Research Center of Machine Learning and Industry Intelligence, Ministry of Education, China · Institute of Data Science, National University of Singapore

Causal Reasoning in Language Models LLM Evaluation LLM Interpretability

Apr 13, 2026

CausalGaze: Unveiling Hallucinations via Counterfactual Graph Intervention in Large Language Models

Despite the groundbreaking advancements made by large language models (LLMs), hallucination remains a critical bottleneck for their deployment in high-stakes domains. Existing classification-based methods mainly rely on static and passive signals from internal states, which often captures the noise and spurious correlations, while overlooking the underlying causal mechanisms. To address this limitation, we shift the paradigm from passive observation to active intervention by introducing CausalGaze, a novel hallucination detection framework based on structural causal models (SCMs). CausalGaze models LLMs' internal states as dynamic causal graphs and employs counterfactual interventions to disentangle causal reasoning paths from incidental noise, thereby enhancing model interpretability. Extensive experiments across four datasets and three widely used LLMs demonstrate the effectiveness of CausalGaze, especially achieving 3.3% improvement in AUROC on the TruthfulQA dataset compared to state-of-the-art baselines.

Linggang Kong, Lei Wu, Yunlong Zhang +4

College of Electronic Engineering, National University of Defense Technology · Anhui Province Key Laboratory of Cyberspace Security Situation Awareness and Evaluation · Institute of Computer Application, China Academy of Engineering Physics

LLM Hallucination Detection Hallucination in Language Models Causal Interventions in Language Models LLM Interpretability

Apr 12, 2026

Attention Sinks as Internal Signals for Hallucination Detection in Large Language Models

Large language models frequently exhibit hallucinations: fluent and confident outputs that are factually incorrect or unsupported by the input context. While recent hallucination detection methods have explored various features derived from attention maps, the underlying mechanisms they exploit remain poorly understood. In this work, we propose SinkProbe, a hallucination detection method grounded in the observation that hallucinations are deeply entangled with attention sinks - tokens that accumulate disproportionate attention mass during generation - indicating a transition from distributed, input-grounded attention to compressed, prior-dominated computation. Importantly, although sink scores are computed solely from attention maps, we find that the classifier preferentially relies on sinks whose associated value vectors have large norms. Moreover, we show that previous methods implicitly depend on attention sinks by establishing their mathematical relationship to sink scores. Our findings yield a novel hallucination detection method grounded in theory that produces state-of-the-art results across popular datasets and LLMs.

Jakub Binkowski, Kamil Adamczewski, Tomasz Kajdanowicz

Department of Artificial Intelligence, Wroclaw University of Science and Technology

LLM Hallucination Detection Attention Sinks LLM Interpretability

Apr 10, 2026

Steered LLM Activations are Non-Surjective

Activation steering is a popular white-box control technique that modifies model activations to elicit an abstract change in its behavior. It has also become a standard tool in interpretability (e.g., probing truthfulness, or translating activations into human-readable explanations) and safety research (e.g., jailbreakability). However, it is unclear whether steered behavior is realizable by any textual prompt. In this work, we cast this question as a surjectivity problem: for a fixed model, does every steered activation admit a preimage under the model's natural forward pass? Under practical assumptions, we prove that activation steering pushes the residual stream off the manifold of states reachable from discrete prompts. Almost surely, no prompt can reproduce the same internal behavior induced by steering. We also illustrate this finding empirically across three widely used LLMs. Our results establish a formal separation between white-box steerability and black-box prompting. We therefore caution against interpreting the ease and success of activation steering as evidence of prompt-based interpretability or vulnerability, and argue for evaluation protocols that explicitly decouple white-box and black-box interventions.

Aayush Mishra, Daniel Khashabi, Anqi Liu

Johns Hopkins University

Activation Steering LLM Interpretability

Apr 10, 2026

The Myth of Expert Specialization in MoEs: Why Routing Reflects Geometry, Not Necessarily Domain Expertise

Mixture of Experts (MoEs) are now ubiquitous in large language models, yet the mechanisms behind their "expert specialization" remain poorly understood. We show that, since MoE routers are linear maps, hidden state similarity is both necessary and sufficient to explain expert usage similarity, and specialization is therefore an emergent property of the representation space, not of the routing architecture itself. We confirm this at both token and sequence level across five pre-trained models. We additionally prove that load-balancing loss suppresses shared hidden state directions to maintain routing diversity, which might provide a theoretical explanation for specialization collapse under less diverse data, e.g. small batch. Despite this clean mechanistic account, we find that specialization patterns in pre-trained MoEs resist human interpretation: expert overlap between different models answering the same question is no higher than between entirely different questions (

\sim

60%); prompt-level routing does not predict rollout-level routing; and deeper layers exhibit near-identical expert activation across semantically unrelated inputs, especially in reasoning models. We conclude that, while the efficiency perspective of MoEs is well understood, understanding expert specialization is at least as hard as understanding LLM hidden state geometry, a long-standing open problem in the literature.

Xi Wang, Soufiane Hayou, Eric Nalisnick

Johns Hopkins University

Mixture-of-Experts Expert Specialization Mixture of Experts Mixture-of-Experts Models Expert Routing LLM Interpretability Representation Geometry in Language Models

Apr 10, 2026

Single-Position Intervention Fails: Distributed Output Templates Drive In-Context Learning

Understanding how large language models encode task identity from few-shot demonstrations is a central open problem in mechanistic interpretability. Prior work uses linear probing to localize task representations, reporting high classification accuracy at specific layers. We reveal a striking dissociation: probing accuracy completely fails to predict causal importance. Single-position activation intervention achieves 0% task transfer across all 28 layers of Llama-3.2-3B-despite 100% probing accuracy at those same positions. This null result is itself a key finding, demonstrating that task encoding is fundamentally distributed. Multi-position intervention-replacing activations at all demonstration output tokens simultaneously-achieves up to 96% transfer (N=50, 95% CI: [87%, 99%]) at layer 8, pinpointing for the first time the causal locus of ICL task identity. We establish the generality of these findings across four models spanning three architecture families (LLaMA, Qwen, Gemma), discovering a universal intervention window at ~30% network depth. Causal tracing uncovers an asymmetric architecture: the query position is strictly necessary (53-100% disruption) while no individual demonstration position is necessary (0% disruption)-resolving a key ambiguity in prior accounts. Crucially, transfer depends on internal representation compatibility, not surface similarity (r=-0.05 vs r=0.31), ruling out trivial explanations. These results establish the distributed template hypothesis: ICL task identity is encoded as output format templates distributed across demonstration tokens, fundamentally reshaping our understanding of how in-context learning operates.

Bryan Cheng, Jasper Zhang

William A. Shine Great Neck South High School

In-Context Learning LLM Interpretability Causal Interventions in Language Models

Apr 9, 2026

What Drives Representation Steering? A Mechanistic Case Study on Steering Refusal

Applying steering vectors to large language models (LLMs) is an efficient and effective model alignment technique, but we lack an interpretable explanation for how it works--specifically, what internal mechanisms steering vectors affect and how this results in different model outputs. To investigate the causal mechanisms underlying the effectiveness of steering vectors, we conduct a comprehensive case study on refusal. We propose a multi-token activation patching framework and discover that different steering methodologies leverage functionally interchangeable circuits when applied at the same layer. These circuits reveal that steering vectors primarily interact with the attention mechanism through the OV circuit while largely ignoring the QK circuit. Freezing all attention scores during steering drops performance by only 8.83% across three model families. A mathematical decomposition of the steered OV circuit further reveals semantically interpretable concepts, even in cases where the steering vector itself does not. Leveraging the activation patching results, we show that steering vectors can be sparsified by up to 85-96% while retaining most performance, and that different steering methodologies agree on a subset of important dimensions.

Stephen Cheng, Sarah Wiegreffe, Dinesh Manocha

University of Maryland, College Park

Language Model Steering LLM Interpretability Representation Steering Attention Mechanisms Circuits in Language Models Refusal Behavior in Language Models

Apr 9, 2026

Distributed Multi-Layer Editing for Rule-Level Knowledge in Large Language Models

Large language models store not only isolated facts but also rules that support reasoning across symbolic expressions, natural language explanations, and concrete instances. Yet most model editing methods are built for fact-level knowledge, assuming that a target edit can be achieved through a localized intervention. This assumption does not hold for rule-level knowledge, where a single rule must remain consistent across multiple interdependent forms. We investigate this problem through a mechanistic study of rule-level knowledge editing. To support this study, we extend the RuleEdit benchmark from 80 to 200 manually verified rules spanning mathematics and physics. Fine-grained causal tracing reveals a form-specific organization of rule knowledge in transformer layers: formulas and descriptions are concentrated in earlier layers, while instances are more associated with middle layers. These results suggest that rule knowledge is not uniformly localized, and therefore cannot be reliably edited by a single-layer or contiguous-block intervention. Based on this insight, we propose Distributed Multi-Layer Editing (DMLE), which applies a shared early-layer update to formulas and descriptions and a separate middle-layer update to instances. While remaining competitive on standard editing metrics, DMLE achieves substantially stronger rule-level editing performance. On average, it improves instance portability and rule understanding by 13.91 and 50.19 percentage points, respectively, over the strongest baseline across GPT-J-6B, Qwen2.5-7B, Qwen2-7B, and LLaMA-3-8B. The code is available at https://github.com/Pepper66/DMLE.

Yating Wang, Wenting Zhao, Yaqi Zhao +3

School of Software, Shandong University, Jinan, China · Salesforce AI Research

Knowledge Editing Language Model Editing Causal Interventions in Language Models LLM Interpretability

Apr 9, 2026

Linear Representations of Hierarchical Concepts in Language Models

We investigate how and to what extent hierarchical relations (e.g., Japan

\subset

Eastern Asia

\subset

Asia) are encoded in the internal representations of language models. Building on Linear Relational Concepts, we train linear transformations specific to each hierarchical depth and semantic domain, and characterize representational differences associated with hierarchical relations by comparing these transformations. Going beyond prior work on the representational geometry of hierarchies in LMs, our analysis covers multi-token entities and cross-layer representations. Across multiple domains we learn such transformations and evaluate in-domain generalization to unseen data and cross-domain transfer. Experiments show that, within a domain, hierarchical relations can be linearly recovered from model representations. We then analyze how hierarchical information is encoded in representation space. We find that it is encoded in a relatively low-dimensional subspace and that this subspace tends to be domain-specific. Our main result is that hierarchy representation is highly similar across these domain-specific subspaces. Overall, we find that all models considered in our experiments encode concept hierarchies in the form of highly interpretable linear representations.

Masaki Sakata, Benjamin Heinzerling, Takumi Ito +2

Tohoku University · RIKEN · Langsmith Inc. +2

Hierarchical Representation Learning LLM Interpretability Representation Geometry in Language Models

Apr 9, 2026

Emotion Concepts and their Function in a Large Language Model

Large language models (LLMs) sometimes appear to exhibit emotional reactions. We investigate why this is the case in Claude Sonnet 4.5 and explore implications for alignment-relevant behavior. We find internal representations of emotion concepts, which encode the broad concept of a particular emotion and generalize across contexts and behaviors it might be linked to. These representations track the operative emotion concept at a given token position in a conversation, activating in accordance with that emotion's relevance to processing the present context and predicting upcoming text. Our key finding is that these representations causally influence the LLM's outputs, including Claude's preferences and its rate of exhibiting misaligned behaviors such as reward hacking, blackmail, and sycophancy. We refer to this phenomenon as the LLM exhibiting functional emotions: patterns of expression and behavior modeled after humans under the influence of an emotion, which are mediated by underlying abstract representations of emotion concepts. Functional emotions may work quite differently from human emotions, and do not imply that LLMs have any subjective experience of emotions, but appear to be important for understanding the model's behavior.

Nicholas Sofroniew, Isaac Kauvar, William Saunders +13

Anthropic

Emotion Representation in Language Models LLM Interpretability LLM Alignment

Apr 8, 2026

Reasoning Fails Where Step Flow Breaks

Large reasoning models (LRMs) that generate long chains of thought now perform well on multi-step math, science, and coding tasks. However, their behavior is still unstable and hard to interpret, and existing analysis tools struggle with such long, structured reasoning traces. We introduce Step-Saliency, which pools attention--gradient scores into step-to-step maps along the question--thinking--summary trajectory. Across several models, Step-Saliency reveals two recurring information-flow failures: Shallow Lock-in, where shallow layers over-focus on the current step and barely use earlier context, and Deep Decay, where deep layers gradually lose saliency on the thinking segment and the summary increasingly attends to itself and the last few steps. Motivated by these patterns, we propose StepFlow, a saliency-inspired test-time intervention that adjusts shallow saliency patterns measured by Step-Saliency via Odds-Equal Bridge and adds a small step-level residual in deep layers via Step Momentum Injection. StepFlow improves accuracy on math, science, and coding tasks across multiple LRMs without retraining, indicating that repairing information flow can recover part of their missing reasoning performance.

Xiaoyu Xu, Yulan Pan, Xiaosong Yuan +4

Shanghai Jiao Tong University · Fuzhou University · Jilin University

LLM Reasoning Large Reasoning Models Inference-Time Intervention LLM Interpretability

Apr 8, 2026

Latent Structure of Affective Representations in Large Language Models

The geometric structure of latent representations in large language models (LLMs) is an active area of research, driven in part by its implications for model transparency and AI safety. Existing literature has focused mainly on general geometric and topological properties of the learnt representations, but due to a lack of ground-truth latent geometry, validating the findings of such approaches is challenging. Emotion processing provides an intriguing testbed for probing representational geometry, as emotions exhibit both categorical organization and continuous affective dimensions, which are well-established in the psychology literature. Moreover, understanding such representations carries safety relevance. In this work, we investigate the latent structure of affective representations in LLMs using geometric data analysis tools. We present three main findings. First, we show that LLMs learn coherent latent representations of affective emotions that align with widely used valence--arousal models from psychology. Second, we find that these representations exhibit nonlinear geometric structure that can nonetheless be well-approximated linearly, providing empirical support for the linear representation hypothesis commonly assumed in model transparency methods. Third, we demonstrate that the learned latent representation space can be leveraged to quantify uncertainty in emotion processing tasks. Our findings suggest that LLMs acquire affective representations with geometric structure paralleling established models of human emotion, with practical implications for model interpretability and safety.

Benjamin J. Choi, Melanie Weber

Harvard University

Emotion Representation in Language Models Representation Geometry in Language Models LLM Interpretability

Apr 7, 2026

Distributed Interpretability and Control for Large Language Models

Large language models that require multiple GPU cards to host are usually the most capable models. It is necessary to understand and steer these models, but the current technologies do not support the interpretability and steering of these models in the multi-GPU setting as well as the single-GPU setting. We present a practical implementation of activation-level interpretability (logit lens) and steering (steering vector) that scales up to multi-GPU language models. Our system implements design choices that reduce the activation memory by up to 7x and increase the throughput by up to 41x compared to a baseline on identical hardware. We demonstrate the method across LLaMA-3.1 (8B, 70B) and Qwen-3 (4B, 14B, 32B), sustaining 20-100 tokens/s while collecting full layer-wise activation trajectories for sequences of 1,500 tokens. Using label-position steering vectors injected post-LayerNorm, we show controllable, monotonic shifts in model outputs with a mean steerability slope of 0.702 across evaluated datasets, without fine-tuning or additional forward passes. We release detailed benchmarks, ablations, and a reproducible instrumentation recipe to enable practical interpretability and real-time behavioral control for frontier LLMs at https://github.com/Devdesai1901/LogitLense.

Dev Arpan Desai, Shaoyi Huang, Zining Zhu

LLM Interpretability Activation Steering

Apr 7, 2026

Attention Flows: Tracing LLM Conceptual Engagement via Story Summaries

Although LLM context lengths have grown, there is evidence that their ability to integrate information across long-form texts has not kept pace. We evaluate one such understanding task: generating summaries of novels. When human authors of summaries compress a story, they reveal what they consider narratively important. Therefore, by comparing human and LLM-authored summaries, we can assess whether models mirror human patterns of conceptual engagement with texts. To measure conceptual engagement, we align sentences from 150 human-written novel summaries with the specific chapters they reference. We demonstrate the difficulty of this alignment task, which indicates the complexity of summarization as a task. We then generate and align additional summaries by nine state-of-the-art LLMs for each of the 150 reference texts. Comparing the human and model-authored summaries, we find both stylistic differences between the texts and differences in how humans and LLMs distribute their focus throughout a narrative, with models emphasizing the ends of texts. Comparing human narrative engagement with model attention mechanisms suggests explanations for degraded narrative comprehension and targets for future development. We release our dataset to support future research.

Rebecca M. M. Hicke, Sil Hamilton, David Mimno +1

Information Science Cornell University · School of Communication and Culture Aarhus University

LLM Evaluation Long-Context Language Model Evaluation Text Summarization LLM Interpretability Summarization Evaluation

Apr 7, 2026

The Illusion of Superposition? A Principled Analysis of Latent Thinking in Language Models

Latent reasoning via continuous chain-of-thoughts (Latent CoT) has emerged as a promising alternative to discrete CoT reasoning. Operating in continuous space increases expressivity and has been hypothesized to enable superposition: the ability to maintain multiple candidate solutions simultaneously within a single representation. Despite theoretical arguments, it remains unclear whether language models actually leverage superposition when reasoning using latent CoTs. We investigate this question across three regimes: a training-free regime that constructs latent thoughts as convex combinations of token embeddings, a fine-tuned regime where a base model is adapted to produce latent thoughts, and a from-scratch regime where a model is trained entirely with latent thoughts to solve a given task. Using Logit Lens and entity-level probing to analyze internal representations, we find that only models trained from scratch exhibit signs of using superposition. In the training-free and fine-tuned regimes, we find that the superposition either collapses or is not used at all, with models discovering shortcut solutions instead. We argue that this is due to two complementary phenomena: i) pretraining on natural language data biases models to commit to a token in the last layers ii) capacity has a huge effect on which solutions a model favors. Together, our results offer a unified explanation for when and why superposition arises in continuous chain-of-thought reasoning, and identify the conditions under which it collapses.

Michael Rizvi-Martel, Guillaume Rabusseau, Marius Mosbach

Mila – Quebec AI Institute · Universit´e de Montr´eal · McGill University

Continuous Latent Reasoning LLM Interpretability Representation Geometry in Language Models

Apr 7, 2026

MechELK: A Mechanistic Interpretability Framework for Eliciting Latent Knowledge in Large Language Models

Large language models (LLMs) frequently encode factual and reasoning knowledge in their internal representations that is not faithfully reflected in their surface-level outputs -- a phenomenon known as latent knowledge. Existing approaches to eliciting latent knowledge, such as Contrastive Consistency Search (CCS), rely on contrastive activation patterns and struggle with complex multi-step reasoning tasks, while mechanistic interpretability tools have primarily been used to understand model behavior rather than to extract hidden knowledge. We present MechELK, a unified three-stage framework that bridges mechanistic interpretability and latent knowledge elicitation. MechELK operates through: (1) Locate -- using Sparse Autoencoder (SAE) feature analysis and activation patching to identify knowledge-bearing representations; (2) Verify -- employing causal probing to distinguish genuine latent knowledge from spurious correlations; and (3) Elicit -- applying representation engineering to surface hidden knowledge without modifying model weights. Evaluated on TruthfulQA, a curated Deceptive Alignment benchmark, and the Quirky LM dataset, MechELK achieves an average elicitation accuracy of 84.7%, outperforming CCS by 6.2% and direct linear probing by 9.1%. Crucially, MechELK successfully identifies latent knowledge in 78.3% of cases where the model's surface output is incorrect or evasive, demonstrating its utility for AI safety applications including deceptive alignment detection.

Ji-jun Park, Soo-joon Choi, Jiwon Jeong +2

Dongguk University

LLM Interpretability Language Model Probing Mechanistic Interpretability

Apr 7, 2026

How LLMs Follow Instructions: Skillful Coordination, Not a Universal Mechanism

Instruction tuning is commonly assumed to endow language models with a domain-general ability to follow instructions, yet the underlying mechanism remains poorly understood. Does instruction-following rely on a universal mechanism or compositional skill deployment? We investigate this through diagnostic probing across nine diverse tasks in three instruction-tuned models. Our analysis provides converging evidence against a universal mechanism. First, general probes trained across all tasks show selective rather than uniform deficits relative to task-specific specialists, indicating that representational sharing is partial and structured rather than global. Second, cross-task transfer is weak and clustered by skill similarity. Third, causal ablation reveals sparse asymmetric dependencies rather than shared representations. Tasks also stratify by complexity across layers, with structural constraints emerging early and semantic tasks emerging late. Finally, temporal analysis shows that the constraint signal becomes decodable only once generation is under way, and remains so throughout the response. These findings indicate that instruction-following is better characterized as skillful coordination of diverse linguistic capabilities rather than deployment of a single abstract constraint-checking process.

Elisabetta Rocchetti, Alfio Ferrara

Department of Computer Science, Università degli Studi di Milano

LLM Interpretability Instruction Following Instruction Tuning

Apr 7, 2026

Disentangling MLP Neuron Weights in Vocabulary Space

Interpreting the information encoded in language model weights remains a fundamental challenge in mechanistic interpretability. In this work, we introduce ROTATE (Rotation-Optimized Token Alignment in weighT spacE), a data-free method requiring no forward passes that disentangles MLP neurons directly in weight space. Our approach relies on a key statistical observation: neurons that encode coherent, monosemantic concepts exhibit high kurtosis when projected onto the model's vocabulary. By optimizing rotations of neuron weights to maximize their vocabulary-space kurtosis, our method recovers sparse, interpretable directions which we name vocabulary channels. Experiments on Llama-3.1-8B-Instruct and Gemma-2-2B-it demonstrate that ROTATE consistently recovers vocabulary channels that are faithful to the neuron's behavior; ablating individual channels selectively disables corresponding input activations or the promotion of specific concepts. Moreover, aggregating channel-level descriptions yields comprehensive neuron descriptions that outperform optimized activation-based baselines by 2-3x in head-to-head comparisons. By providing a data-free decomposition of neuron weights, ROTATE offers a scalable, fine-grained building block for interpreting language models.

Asaf Avrahamy, Yoav Gur-Arieh, Mor Geva

Blavatnik School of Computer Science and AI, Tel Aviv University

LLM Interpretability Transformer Interpretability Mechanistic Interpretability

Apr 7, 2026

What are They Thinking? Delineation, Probing, and Tracking of Concepts in LLMs

As the influence of LLMs expands, it is imperative to gain insight into their decisions. One way to do that is to develop probes that detect the presence or absence of a broad set of high-level abstract concepts within the embeddings computed in an LLM - which is what we might say a model is ``thinking" about. Such probes should be low-cost and easily applicable to any LLM, so that monitoring for many concepts is possible during normal operation. In this paper, we take the first steps towards developing the capability of creating many such probes by defining and executing examples of the key tasks needed: first, the careful delineation of a high-level abstract concept through the creation of a dataset with the concept both present and then absent. Then, the training and testing of a set of linear probes to detect the concept on any layer of an LLM, including an exploration of the complexity of the probe needed. Finally, we show that such probes can track concepts across larger contexts. This is done with four separate concepts and three different LLMs. When this process is scaled to many more concepts, it will create the ability to monitor new models.

Mohamed Abdelwahab, Michelle Yu Collins, Sihan Chen +5

The Edward S. Rogers Sr. Department of Electrical and Computer Engineering University of Toronto

Linear Probing LLM Interpretability Language Model Probing

Apr 7, 2026

Attribution-Driven Explainable Intrusion Detection with Encoder-Based Large Language Models

Software-Defined Networking (SDN) improves network flexibility but also increases the need for reliable and interpretable intrusion detection. Large Language Models (LLMs) have recently been explored for cybersecurity tasks due to their strong representation learning capabilities; however, their lack of transparency limits their practical adoption in security-critical environments. Understanding how LLMs make decisions is therefore essential. This paper presents an attribution-driven analysis of encoder-based LLMs for network intrusion detection using flow-level traffic features. Attribution analysis demonstrates that model decisions are driven by meaningful traffic behavior patterns, improving transparency and trust in transformer-based SDN intrusion detection. These patterns align with established intrusion detection principles, indicating that LLMs learn attack behavior from traffic dynamics. This work demonstrates the value of attribution methods for validating and trusting LLM-based security analysis.

Umesh Biswas, Shafqat Hasan, Syed Mohammed Farhan +2

Computer Science & Engineering, Mississippi State University

Explainable Intrusion Detection LLMs for Cybersecurity Network Intrusion Detection LLM Interpretability Feature Attribution

Apr 6, 2026

Are Latent Reasoning Models Easily Interpretable?

Latent reasoning models (LRMs) have attracted significant research interest due to their low inference cost (relative to explicit reasoning models) and theoretical ability to explore multiple reasoning paths in parallel. However, these benefits come at the cost of reduced interpretability: LRMs are difficult to monitor because they do not reason in natural language. This paper presents an investigation into LRM interpretability by examining two state-of-the-art LRMs. First, we find that latent reasoning tokens are often unnecessary for LRMs' predictions; on logical reasoning datasets, LRMs can almost always produce the same final answers without using latent reasoning at all. This underutilization of reasoning tokens may partially explain why LRMs do not consistently outperform explicit reasoning methods and raises doubts about the stated role of these tokens in prior work. Second, we demonstrate that when latent reasoning tokens are necessary for performance, we can decode gold reasoning traces up to 65-93% of the time for correctly predicted instances. This suggests LRMs often implement the expected solution rather than an uninterpretable reasoning process. Finally, we present a method to decode a verified natural language reasoning trace from latent tokens without knowing a gold reasoning trace a priori, demonstrating that it is possible to find a verified trace for a majority of correct predictions but only a minority of incorrect predictions. Our findings highlight that current LRMs largely encode interpretable processes, and interpretability itself can be a signal of prediction correctness.

Connor Dilgren, Sarah Wiegreffe

Department of Computer Science, University of Maryland, College Park, MD, USA

Implicit Reasoning in Language Models LLM Reasoning LLM Interpretability