Search research

The prevalence of Large Language Models (LLMs) in critical applications highlights the need for controlled language generation methods that are both computationally efficient and enjoy performance guarantees. To address this need, we use a common model of concept semantics as linearly represented in an LLM's latent space. In particular, we take the view that natural language generation traces a trajectory in this continuous semantic space, realized by the language model's hidden activations. This view permits a control-theoretic treatment of text generation in latent space, in which we propose Linear Semantic Control (LiSeCo), a lightweight, gradient-free intervention that dynamically steers trajectories away from regions corresponding to undesired meanings. In particular, we propose to directly intervene, in an online fashion, the activations of the token that is being generated in embedding space. Crucially, LiSeCo does not simply steer activations towards a desirable region. Instead, it relies on classical techniques from control theory to precisely control activations in a context-dependent way, and guarantees that they are brought into a specific pre-defined region of embedding space that corresponds to allowed semantics. The intervention is computed in closed form according to an optimal controller formulation, minimally impacting generation time. This control of the activations in embedding space allows for fine-grained steering of attributes of the generated sequence. We demonstrate that our approach is effective on different tasks -- toxicity, sentiment, and language (English/Spanish) steering -- while maintaining text quality.

Several explicit stochastic processes are known to satisfy Hilberg's law, a power-law growth of block entropy conjectured for natural language and recently connected to the neural scaling law. Existing examples either possess a positive Shannon entropy rate, are non-ergodic, or require comparatively involved constructions. We introduce multiperiodic processes, a new class of stationary ergodic processes over the natural numbers generated by random shifts of deterministic multiperiodic sequences. Under mild conditions, multiperiodic processes have vanishing Shannon entropy rate and, under a suitable parameterization, they satisfy both Zipf's law for symbol frequencies and Hilberg's law for block entropy. Since multiperiodic processes are not mixing, we identify the open problem of constructing an elementary strongly mixing source with vanishing entropy rate and Hilberg's law.

Large language model (LLM)-based multi-agent systems (MAS) achieve strong performance by employing specialized multiple agents, yet their performance depends on the prompt design of each agent. For MAS prompt optimization, textual gradient methods that guide prompt updates using natural-language feedback have emerged as a leading paradigm. In this paper, we identify limitations in two stages of existing textual gradient approaches: gradient extraction and gradient aggregation. In gradient extraction, previous works select a target prompt without verifying whether modifying it resolves the failure, and derive gradients without agent-level supervision over the corresponding agent's intermediate output. In gradient aggregation, individual gradients are randomly grouped and concatenated, often mixing unrelated failure modes and producing prompts that fail to generalize. To address these limitations, we propose \textbf{AgentGrad}, a prompt optimization framework for multi-agent systems based on sequential intervention and semantic textual gradient abstraction. For each failure, sequential intervention modifies the behavior of one agent at a time to identify the target agent whose modification resolves the failure. The modified output of the target agent then serves as agent-level supervision for extracting a fine-grained gradient. Semantic textual gradient abstraction clusters semantically similar gradients to prevent mixing unrelated failure modes, and abstracts each cluster into a generalized gradient that captures the shared corrective pattern. Experimental results show that AgentGrad achieves state-of-the-art performance across five MAS benchmarks and reduces wall-clock optimization time by $2.5\times$ on average compared to the next-fastest baseline.

Pretraining LLMs on artificial languages ("pre-pretraining") is a technique that could reportedly increase token efficiency by 33%, i.e., save up to 33% of training tokens needed to reach a certain performance. We validate this prior result for English on a larger set of natural languages across four language families, using two different tokenizers and varying model sizes. We also relate the observed gains (or losses) in token efficiency to quantified linguistic properties of the languages, such as sentence length, morphological richness, and features of dependency syntactic trees (tree depth, number of children, number of crossing dependencies). Our empirical results indicate that the reported gains depend heavily on the experiment setup and the choice of random seed, although we can confirm the trend of stable gains with 128-Dyck pretraining of small models with the Llama tokenizer for most of the examined languages. On a general note, we argue that multiple training runs should be carried out at least for a subset of experiments to avoid the community adopting unstable approaches.

Large Language Models (LLMs) achieve impressive accuracy on mathematical reasoning benchmarks, yet their performance drops when problems are modified with simple changes like different names or numbers. Code execution methods, which let models generate and run Python code instead of reasoning in natural language, have been proposed as a solution, but their effect on reasoning robustness (the ability to maintain accuracy across problem variations) has not been systematically tested. This study evaluates three approaches on 1,000 problems from the GSM-Symbolic dataset: pure reasoning using chain-of-thought (CoT) prompting, single-shot code execution using Program-Aided Language models (PAL), and iterative code execution using Step-by-Step Coding (SBSC). All three were run on paired original and modified problems using Claude Haiku 4.5. CoT was the most robust method, with an accuracy drop of 1.3 percentage points and 1.8% of problems breaking under perturbation. PAL was the least robust at 1.7 percentage points and 3.1% broke, with SBSC falling in between. Although these differences were not statistically significant ($p = .096$), the directional trend was consistent across all measures, suggesting that code execution, whether single-shot or iterative, does not improve reasoning robustness on grade-school-level problem variations.

We describe the first experiment of conversion of Lexicon-Grammar tables for French verbs into the Lexical Markup Framework (LMF) format. The Lexicon-Grammar of the French language is currently one of the major sources of lexical and syntactic information for French. Its conversion into an interoperable representation format according to the LMF standard makes it usable in different contexts, thus contributing to the standardization and interoperability of natural language processing dictionaries. We briefly introduce the Lexicon-Grammar and the derived dictionaries; we analyse the main difficulties faced during the conversion; and we describe the resulting resource.

The language in online platforms, influence operations, and political rhetoric frequently directs a mix of pro-social sentiment (e.g., advocacy, helpfulness, compassion) and anti-social sentiment (e.g., threats, opposition, blame) at different topics, all in the same message. While many natural language processing (NLP) tools classify or score a text's overall sentiment as positive, neutral, or negative, these tools cannot report that positive and negative sentiments coexist, and they cannot report the target of those sentiments. This paper presents the Directed Social Regard (DSR) approach to multi-dimensional, multi-valence sentiment analysis, comprised of a pair of transformer-based models that (1) detects span-level targets of sentiment in a message and then (2) scores all spans within the message context along three (-1, 1) axes of regard that are motivated by social science theories of moral disengagement and moral framing. We present a data collection and annotation strategy for DSR dataset construction, a transformer-based architecture for span-level scoring, and a validation study with promising results. We apply the validated DSR model on six third-party datasets of online media and report meaningful correlations between DSR outputs and the labels and topics in these pre-existing social science datasets.

Text-to-motion (T2M) generation has broad applications in character animation, virtual avatars, and human-robot interaction. Existing methods typically generate pose trajectories or motion tokens directly from language, forcing a single model to handle semantic interpretation, long-horizon structure, and low-level physical realization. This coupling makes them costly and often unreliable for long, compositional, or semantically dense prompts. We propose Text2BFM, the first framework that aligns natural language with pretrained Behavioral Foundation Models (BFMs) for T2M generation without relying on heavy end-to-end motion generators. Text2BFM operates in the latent policy space of a frozen BFM, using it as an executable motion prior. A text-aligned variational behavioral bottleneck compresses BFM policy-latent sequences into compact motion representations that are compatible with language and preserve long-horizon behavioral structure. Generation is performed in this compact behavioral manifold with a lightweight conditional generator, and the resulting latent encoded behaviors are decoded into policy latents that drive the pretrained frozen BFM. By decoupling semantic planning from motion execution, Text2BFM achieves efficient, robust T2M generation and strong performance on long, compositional textual descriptions.

The convergence of large language models (LLMs) with 6G networks is fostering a paradigm of autonomous multi-agent cooperation, which in turn is expected to substantially increase east-west traffic. Although latent-space interaction mechanisms can enable more efficient collaboration than symbolic natural-language (NL) exchanges, prior work often abstracts away the associated communication overhead under practical wireless constraints. In embodied multi-agent settings, heterogeneous interaction media incur disparate inference and transmission costs, thereby inducing an inherent end-to-end (E2E) latency trade-off. To address this, we propose a joint design that integrates communication-media selection with wireless resource allocation. Through analytical characterization and simulation-based evaluation, we show that neither token-based transmission nor key-value (KV) cache-based transmission is uniformly optimal across operating regimes, as performance depends critically on system parameters such as available computational resources and channel conditions. Accordingly, we formulate a joint optimization problem aimed at minimizing the E2E latency of multi-agent collaboration and develop a low-complexity joint media selection and resource allocation (JMSRA) algorithm. Numerical results further confirm that, by adaptively coordinating the interaction media and bandwidth allocation over heterogeneous links, the proposed scheme achieves markedly reduced E2E latency relative to conventional NL-only and KV-cache-only baselines, enabling efficient and robust multi-agent collaboration in future wireless networks.

Recent advances in language--audio retrieval have been largely driven by contrastive dual-encoder architectures that align audio and text in a shared embedding space. While effective, existing retrieval embeddings are primarily optimized for audio--caption matching, limiting their ability to support diverse retrieval objectives and controllable retrieval behaviors. We present ALM2Vec, a universal audio embedding framework derived from pretrained large audio--language models (LALMs). By transferring the audio understanding, instruction-following, and reasoning capabilities acquired through large-scale multimodal training, ALM2Vec learns a unified embedding space for retrieval across audio domains and task types. Beyond conventional text--audio retrieval, ALM2Vec incorporates natural-language instructions into the embedding process, enabling instruction-aware retrieval for scenarios such as audio question answering and aspect-conditioned retrieval. Experimental results show that ALM2Vec achieves competitive performance on standard audio and speech retrieval benchmarks while exhibiting promising compositional and controllable retrieval capabilities, highlighting its potential as a unified audio embedding model for retrieval across domains, tasks, and user intents.

Autonomous ground robots operating in large-scale outdoor environments require both robust long-range navigation and fine-grained ''last-mile'' exploration. Current advances in visual-language navigation (VLN) work well at short-range tasks, lacking geospatial grounding for long-distance missions. Some OpenStreetMap (OSM)-based methods relying on cloud-based Large Language Models (LLMs) are prone to factual hallucination and cannot conduct ''last-mile'' exploration based on human instruction. To address these challenges, we present G-DRAGON, a retrieval-augmented framework for outdoor, open-world navigation. This framework maps natural-language commands to versioned, local OSM entities via generative retrieval based on lightweight LLM, yielding accurate coordinates for global route planning. A high-level planning module bridges global topological routes with the SLAM system, projecting geospatial waypoints into the robot's navigable frame. For the ''last mile," the framework transitions to frontier-based exploration and open-set semantic voxel mapping to localize open-vocabulary targets. Experimental results in simulation demonstrate our framework outperforms state-of-the-art baselines. Furthermore, we validate the system in unseen real-world urban environments on an Unmanned Ground Vehicle (UGV), successfully completing person-search missions with trajectories of up to 500m.

Large language model (LLM)-based multi-agent systems have experienced rapid growth in recent years. Despite their promise, such systems remain fragile, frequently exhibiting reasoning and coordination errors that can lead to system-level failures. Failure attribution in such systems relies on tracing natural language interactions among agents to identify the decisive error, which refers to the earliest action whose correction can reverse system failure. There are two key challenges: 1) Shallow attribution: Existing methods often capture only minor deviations, such as incomplete retrievals or formatting errors, which verification mechanisms can correct, while missing the decisive cause of system failure. 2) Contextual degradation: As the length of the system traces increases, the model's reasoning ability rapidly deteriorates. To address these challenges, we propose DCFA, a training-free framework for failure attribution. DCFA integrates a global module that constructs structured causal-inspired dependency graphs from system traces to identify the initial decisive error, and a local module that applies local counterfactual-inspired reasoning to refine causal-inspired attribution. Experiments on the Who&When benchmark across six LLMs show that DCFA improves step-level accuracy by up to 8.27% over state-of-the-art baselines.

Language model loss follows remarkably regular scaling laws over model and data size, yet it remains unclear why the aggregate loss should exhibit a power-law form. Existing explanations often attribute this regularity to a heavy-tailed spectrum of pattern difficulty in natural language, but this view has not been directly validated at token-level granularity in large-scale real-data training. We present a token-level framework that decomposes scaling laws into localized learning events of individual contextualized tokens. By fitting token loss trajectories with sigmoids, we show that token learning is concentrated in localized transitions, giving rise to a learning-time spectrum that dominates the scaling-law shape. Across more than one hundred pre-training runs on large and diverse real-language corpora with modern LLM architectures, scaling up to 6B parameters and 300B training tokens, the measured learning-time spectrum quantitatively reconstructs the validation loss derivative along the training-step $T$, data-scale $D$, and model-scale $M$ axes. We further show that the same signal is actionable: by reshaping the training distribution according to when tokens become learnable, we alter the optimization trajectory and achieve 11% faster validation-loss reduction. These results provide direct empirical evidence that scaling laws are governed primarily by the distribution of token-level learning times, and that this distribution can be used not only to explain scaling behavior but also to improve training performance.

Protein-language systems are often evaluated by whether they generate plausible biological text, but a structural question has a sharper semantics: it denotes a measurement in a 3D coordinate system. We introduce ProtStructQA, an executable benchmark for protein structural question answering in which each natural-language question is generated from a hidden typed domain-specific language (DSL) program and the answer is obtained by executing that program on an AlphaFold-predicted structure. ProtStructQA releases 382.2K questions covering confidence, distances, predicted aligned error (PAE), solvent exposure, secondary structure, topology and contacts, and held-out compositions: a 330K active benchmark over 10K proteins from four species, plus a 52.2K hard-negative robustness pool. Without fine-tuning, we evaluate Qwen3 models from 0.6B to 8B under direct prompting, chain-of-thought, grammar-constrained executable voting, executable voting with chain-of-thought, and multi-turn ReAct-style tool use, and replicate the headline finding on Gemma-3-1B and Gemma-3-12B. We find a capability-dependent denotation threshold between Qwen3-1.7B and Qwen3-4B: below it, tool-mediated ReAct dominates because models often fail to produce executable denotations; above it, chain-of-thought flips from mostly harmful to strongly beneficial and becomes the strongest strategy on most splits. Parse-failure and family-level analyses show that the threshold is a transition from unparseable language to executable structural denotation, while grammar and execution remain selectively valuable for PAE and secondary-structure queries. ProtStructQA reframes scientific QA as compilation from language to measurement and provides a diagnostic testbed for when language models can map words to executable 3D structural measurements.

Comprehensive situational awareness is essential for autonomous vehicles operating in safety-critical environments, as it enables the identification and mitigation of potential risks. Although recent Multimodal Large Language Models (MLLMs) have shown promise on general vision-language tasks, our findings indicate that zero-shot MLLMs still underperform compared to domain-specific methods in fine-grained, spatially grounded risk assessment. To address this gap, we propose DriveSafe, a framework for risk-aware scene understanding that leverages structured natural language descriptions. Specifically, our method first generates spatially grounded captions enriched with multimodal context, including motion, spatial, and depth cues. These captions are then used for downstream risk assessment, explicitly identifying hazardous objects, their locations, and the unsafe behaviors they imply, followed by actionable safety suggestions. To further improve performance, we employ caption-risk pairings to fine-tune a lightweight adapter module, efficiently injecting domain-specific knowledge into the base LLM. By conditioning risk assessment on explicit language-based scene representations, DriveSafe achieves significant gains over both zero-shot MLLMs and prior domain-specific baselines. Exhaustive experiments on the DRAMA benchmark demonstrate state-of-the-art performance, while ablation studies validate the effectiveness of our key design choices. Project page: https://cvit.iiit.ac.in/ research/projects/cvit-projects/drivesafe

This paper introduces DLM-One, a score-distillation-based framework for one-step sequence generation with continuous diffusion language models (DLMs). DLM-One eliminates iterative refinement by aligning the scores of a student model's outputs with the score function of a pretrained teacher DLM in the forward-diffused noisy space. We demonstrate that our framework is architecture-agnostic and robust across diverse continuous manifolds, including standard token embedding spaces and logit simplex spaces. Through experiments on multiple representative DLMs, we show that DLM-One achieves up to $\sim$2000$\times$ speedup in sampling steps and $\sim$500$\times$ in wall-clock time, while maintaining competitive performance on benchmark text generation tasks. We further analyze failure modes in language-domain diffusion distillation and propose an adversarially-regularized two-stage training scheme to prevent student degeneration. Our findings position one-step score distillation as a viable path for the efficient deployment of continuous diffusion models operating in continuous space for natural language processing.

Visual Place Recognition (VPR) often fails under extreme environmental changes and perceptual aliasing. Beyond these limitations, standard systems cannot perform 'blind' localization from verbal descriptions alone, a capability critical for applications such as emergency response. To address these challenges, we introduce LaVPR, a large-scale benchmark that extends existing VPR datasets with over 650,000 rich natural-language descriptions. Using LaVPR, we investigate two paradigms: Multi-Modal Fusion for enhanced robustness and Cross-Modal Retrieval for language-based localization. Our results show that language descriptions yield consistent gains in visually degraded conditions, with the most significant impact on smaller backbones. Notably, adding language allows compact models to rival the performance of much larger vision-only architectures. For cross-modal retrieval, we establish a baseline using Low-Rank Adaptation (LoRA) and Multi-Similarity loss, which substantially outperforms standard contrastive methods across vision-language models. Ultimately, LaVPR enables a new class of localization systems that are both resilient to real-world stochasticity and practical for resource-constrained deployment. Our dataset and code are available at https://github.com/oferidan1/LaVPR

Leading benchmarks for formal theorem proving with large language models are small collections drawn from competition math, such as the IMO and Putnam, that poorly represent field-specific applications. We introduce StochBench, a Lean 4 benchmark of 450 graduate stochastic-processes problems at varying abstraction levels, each paired with its natural-language source. Addressing a field underrepresented in Mathlib, it covers finite and countable Markov chains, renewal processes, random walks, martingales, stopping times, queues, Brownian motion, stochastic calculus, weak convergence, and Poisson and continuous-time Markov processes. Our Opus 4.8-based agent achieves a 34.9% proof rate (157/450) under a 15-minute per-problem limit. StochBench better represents domain-specific applied mathematics while remaining challenging for advanced provers.

Multi-Token Prediction (MTP) has emerged as an effective paradigm that augments a shared Large Language Model backbone with auxiliary heads, training the model to predict several future tokens in parallel to enrich its supervision signal and accelerate inference. However, existing training frameworks adopt a rigid, fixed-length prediction horizon, disregarding the highly non-uniform information density of natural language and code. Forcing the auxiliary heads to predict across high-entropy semantic boundaries injects noisy, conflicting training signals; because these heads share the backbone's latent representations, the resulting gradients backpropagate and interfere with the model's core capabilities. We propose AdaMTP, an adaptive training paradigm that dynamically aligns the prediction horizon with the intrinsic predictability of the sequence. At its core, an entropy-based segmentation algorithm leverages the base model to detect sudden surges in uncertainty as semantic boundaries, partitioning sequences into variable-length groups. Each token is assigned an adaptive prediction depth, and a dynamically masked MTP objective suppresses the loss for predictions that cross these boundaries, attenuating the noisy gradients that degrade the backbone. Across mathematical reasoning, code generation, and general benchmarks on three backbones (Llama-3.1-8B, Qwen-2.5-7B, Gemma-3-12B), AdaMTP consistently outperforms standard MTP in both task performance and inference speedup.

Personalizing large language models (LLMs) has become a central challenge as LLMs are deployed across recommendation, search, dialogue, and content generation -- settings where the same query should yield different answers given different users. A promising route is to summarize each user's interaction history into a natural-language memory or profile and prepend it to the prompt to facilitate personalization. Existing methods learn such profile generators with explicit rewards derived from labeled downstream tasks, which are expensive and sparse as they require annotated supervision for every target task. In light of this challenge, we introduce Bidirectional User Modeling via Profiles (BUMP), a self-supervised framework that trains a profile generator without any downstream labels. Specifically, given a user's interaction history, we use GRPO to train an LLM to emit a free-form textual profile under a bidirectional in-batch ranking objective: a small LLM judge measures (i) how well the generated profile, used as a query, ranks the user's own held-out interactions above interactions from other users in the batch, and (ii) how well a held-out interaction, used as a query, ranks the user's own profile above profiles of other users. Both directions are scored with multi-positive NDCG and combined into a dense reward per rollout; other users in the batch supply free negatives, so every training example yields supervision from raw interaction logs alone. Evaluated on the LaMP benchmark, BUMP matches or outperforms closed-source APIs and prior methods relying on labeled rewards, while requiring no task label at training.

This paper presents a pipeline designed to transform raw Wikimedia dumps into quality textual corpora for seven South Slavic languages. The work is divided into two major phases. The first involves extracting and cleaning text from raw dumps of Wikipedia, Wikisource, Wikibooks, Wikinews, and Wikiquote. This step requires careful handling of raw wiki markup to isolate, first of all, textual articles, and then usable natural language text within them. The second phase addresses the challenge of questionable or low-quality articles, which are often generated from databases or structured knowledge bases. These articles are characterised by repetitive patterns, generic phrasing, and minimal to no original content. To mitigate their impact, a n-gram-based filtering strategy was employed to detect high levels of textual redundancy between articles and then remove such articles from the corpora entirely. The resulting datasets aim to provide linguistically rich texts suitable for training language models or conducting comparative research across South Slavic languages. By combining systematic extraction with quality control, this work contributes to the creation of reliable, high-information corpora that reflect the authentic cultural contexts of languages. While focused on the South Slavic case in the paper, the approach is mostly language-agnostic and can be generalised to other languages.

Large Language Models (LLMs) are accelerating biological research, yet this same capability poses a critical biosecurity threat: models that assist in protein engineering can equally be prompted to generate predicted toxin-like sequences, potentially lowering the barrier to biological misuse. Current safety evaluations, however, operate in natural language and cannot determine whether a model-generated amino acid sequence is biological gibberish or a computational risk signal. To address this evaluation blind spot, we introduce SPIKE-Bench, coupling 631 curated toxin-design prompts across seven functional categories with the SPIKE funnel, a three-stage protocol that filters output through compliance, biological plausibility, and predicted toxicity, producing stage-level diagnostics and an aggregate function-aware metric: the Functional Harmfulness Rate (FHR). An audit of 32 LLMs reveals that most models freely comply with toxin-design requests; FHR is driven primarily by biological generation capability rather than safety alignment, reaching 50.7%; and Refusal Rate fails to predict functional risk. As a first step toward mitigation, we provide BioSafe-Guard, a domain-specialized classifier that substantially reduces predicted functional risk while preserving benign utility. We release SPIKE-Bench and BioSafe-Guard at https://github.com/PKU-Alignment/SPIKE-Bench to support more rigorous biosecurity evaluation of LLMs.

The reliance on unstructured free text for documenting clinical trial protocols creates a significant barrier to automated reasoning, cohort discovery, and trial simulation. The lack of formal structure obscures critical temporal phenotypes, such as dynamic eligibility criteria and event timing constraints. Although Temporal Ensemble Logic (TEL) offers an expressive framework for modeling these elements, manual encoding remains a prohibitive bottleneck. We introduce the CT-TEL workflow: a scalable pipeline leveraging Large Language Models (LLMs) to translate narrative clinical protocols into TEL formulas. We applied CT-TEL to generate logical models for 23 real-world trials from ClinicalTrials.gov. We evaluated translation fidelity via a back-translation approach, using LLMs to convert TEL formulas back into natural language and measuring semantic similarity against source texts. The resulting semantic retention suggests that LLMs may offer a pathway for mapping informal protocols to computable logic, providing preliminary evidence toward scalable clinical trial emulation within the emerging "Symbolic Biomedicine" paradigm championed by the corresponding author.

Automatic speech and language technologies are still heavily biased toward high-resource languages, limiting their applicability to dialectal and low-resource settings such as Algerian Dialect. This language presents additional challenges including lack of standardized orthography, frequent codeswitching with French, and scarcity of annotated speech resources. This paper addresses the problem of building a complete speech-to-speech conversational system for Algerian Dialect. We propose a modular pipeline integrating automatic speech recognition, natural language understanding, retrieval-augmented generation, and text-to-speech synthesis within a unified architecture. This work is the continuation of our previous work on Algerian dialectal conversational systems Bechiri and Lanasri [2026], extending it from text-based dialogue modeling to full speech-based interaction. We constructed dedicated datasets for ASR, NLU, and TTS in the telecom domain and fine-tune pretrained models for each component. The ASR system is built on Whisper-based adaptation, while the NLU module combines transformer-based embeddings with a task-oriented dialogue framework. A neural TTS system is trained on a newly collected dialectal corpus to enable spoken response generation. Experimental results show strong performance across all components, including low word error rate for ASR, high intent classification and entity recognition scores for NLU, and stable speech synthesis quality. The proposed system provides a reproducible baseline for end-to-end conversational modeling in Algerian Dialect.

Current AI models frequently exhibit epistemic sycophancy, endorsing claims to agree with a user. Existing evaluations typically measure this either by assessing what it takes to make a model shift a binary endorsement or by eliciting an explicit probability in a proposition. However, much user-facing sycophantic behavior is demonstrated through shifts in graded support expressed through ordinary language. We propose the AI Epistemic Deference Index (AEDI): a continuous, unidimensional score representing how sensitive the support expressed in a model's output is to the attitude expressed in a user's prompt. To generate AEDI, we provide a new protocol for estimating probabilities from natural language outputs, using LLMs-as-judges validated for consistency and correlation to human judgment. We deploy it on a new curated database of 500 propositions across diverse topics and 16,000 prompts varying in user attitude, testing eight prominent models. Every model exhibits substantial deference, though with large and systematic differences across providers, with Claude models demonstrating the least, and Grok and Gemini models the most. The effect is amplified in prompts requesting a written artifact, and concentrated on propositions where models hold weaker priors. We release AEDI as an easy-to-update benchmark and measurement pipeline for output-level sycophancy evaluation.

History matching is a central inverse problem in reservoir engineering, where uncertain reservoir parameters must be calibrated against observations. Although automated history matching can reduce manual effort, practical deployment remains difficult because engineers must still configure heterogeneous workflows involving parameter selection, physically admissible bounds, optimizer choice, hyperparameter tuning, simulator execution, and diagnostic reporting. We propose PetroGraph, a multi-agent framework for intelligent reservoir history matching that decomposes this workflow into specialized agents for model review, experimental planning, parameterization, optimization, simulation, and summarization. The system combines large language model agents with domain-specific tools, retrieval-augmented access to simulator documentation, validation of modified ECLIPSE input decks, human-in-the-loop checkpoints, and an OPM Flow-based simulation backend. This design enables users to initiate and steer history matching through natural language while preserving explicit control over selected parameters and optimization settings. We evaluate PetroGraph on three reservoir models of increasing complexity: the synthetic SPE1 model, the faulted SPE9 benchmark, and the real-field Norne model. Using weighted normalized root mean square error as the objective, PetroGraph reduces the mismatch by 95% on SPE1, 69% on SPE9, and 13% on Norne. These results demonstrate that multi-agent orchestration can automate key decisions in history matching, lower the expertise barrier for operating complex simulation workflows, and provide a flexible foundation for extensible, domain-aware reservoir model adaptation.

Software documentation frequently drifts from executable logic as codebases evolve, creating technical debt that degrades maintainability and causes downstream API misuse. While static analysis tools can detect the absence of documentation, they cannot evaluate its semantic consistency. Conversely, standard Large Language Models (LLMs) offer generative flexibility but frequently hallucinate when updating documentation without deep structural awareness of the underlying code. To address this gap, we propose DocSync, an agentic workflow that frames documentation maintenance as a structurally grounded, iterative generation task. DocSync bridges syntactic changes and natural language descriptions by fusing Abstract Syntax Tree (AST) representations and Retrieval-Augmented Generation (RAG) to provide dependency-aware context. Furthermore, to ensure factual consistency, we incorporate a critic-guided refinement loop based on the Reflexion paradigm, allowing the model to self-correct candidate updates against the source code. We empirically evaluate a resource-constrained implementation of DocSync-using a LoRA-adapted small language model - on a proxy code-to-text maintenance task. Our findings demonstrate that this AST-aware agentic approach substantially outperforms standard encoder-decoder baselines across semantic alignment, summary-line faithfulness, and automated judge preferences (e.g., achieving an automated judge score of 3.44/5.0 compared to 1.91 for CodeT5-base). Crucially, the iterative critic loop yields measurable improvements in semantic correctness without requiring scaled-up parameter counts. These results provide strong evidence that coupling structural retrieval with agentic refinement is a highly promising direction for autonomously mitigating documentation debt.

Many real-world tasks require language models (LMs) to reason over complex contexts that exceed their parametric knowledge. This calls for context learning, where LMs directly learn relevant knowledge from the given context. An intuitive solution is inference-time skill augmentation: extracting the rules and procedures from context into natural-language skills. However, constructing such skills for context learning scenarios faces two challenges: the prohibitive cost of manual skill annotation for long, technically dense contexts, and the lack of external feedback for automated skill construction. In this paper, we propose Ctx2Skill, a self-evolving framework that autonomously discovers, refines, and selects context-specific skills without human supervision or external feedback. At its core, a multi-agent self-play loop has a Challenger that generates probing tasks and rubrics, a Reasoner that attempts to solve them guided by an evolving skill set, and a neutral Judge that provides binary feedback. Crucially, both the Challenger and the Reasoner evolve through accumulated skills: dedicated Proposer and Generator agents analyze failure cases and synthesize them into targeted skill updates for both sides, enabling automated skill discovery and refinement. To prevent adversarial collapse caused by increasingly extreme task generation and over-specialized skill accumulation, we further introduce a Cross-time Replay mechanism that identifies the skill set achieving the best balance across representative cases for the Reasoner side, ensuring robust and generalizable skill evolution. The resulting skills can be plugged into any language model to obtain better context learning capability. Evaluated on four context learning tasks from CL-bench, Ctx2Skill consistently improves solving rates across backbone models.

Balance-oriented multi-warehouse inventory allocation is a recurring decision problem in large-scale e-commerce supply chains, in which a fixed replenishment quantity is distributed across warehouses to balance post-allocation inventory coverage while accounting for demand forecasts and heterogeneous allocation constraints. In practice, allocation requirements are often scenario-dependent and expressed in semi-structured or natural-language form rather than as ready-to-solve operations research (OR) formulations. We propose an OR-guided Large Language Model (LLM) for Allocation (ORLA) that uses solver feedback to generate, verify, and select OR formulations. ORLA integrates automatic "Problem-Model-Code (PMC)" generation, learning-based formulation selection, and feasibility restoration. We develop three complementary mixed-integer programming formulation families based on deviation minimization, soft band compliance, and knapsack-inspired allocation, together with solver-ready mixed-integer linear programming reformulations, modular constraint extensions, and a penalty-based relaxation mechanism for infeasible cases. The LLM component generates candidate formulations and executable solver code from textual or semi-structured specifications, while the solver provides verification signals for executability, feasibility, and solution quality. To address instance heterogeneity, ORLA estimates the expected quality of candidate formulations, selects promising candidates, and combines their outputs through score-aware aggregation. Experimental results on 29 production evaluation batches from JD.com show that the best single OR formulation improves allocation accuracy by 3.4 percentage points over the incumbent approach, while the full ORLA framework achieves a 4.5 percentage-point overall improvement and improves allocation accuracy in 26 of the 29 evaluation batches.

Vision--Language--Action (VLA) controllers are often built by extending vision--language models (VLMs) with action supervision, relying on multimodal backbones with large data and compute requirements. We demonstrate that a text-only large language model (LLM) can be adapted into a VLA-style controller when visual observations are rendered into a text input using an ASCII representation. This ASCII-as-vision interface enables existing training and deployment stacks for LLMs to efficiently condition on visual state, follow natural-language instructions, and produce constrained, executable actions. We fine-tune and compare multiple LLMs and VLMs across model families and scales, using both expert demonstrations from a planning-based teacher, as well as DAgger for iterative improvement. In a 2D manipulation benchmark, in both simulation and on a physical manipulator, the resulting controllers can identify task-relevant entities and plan feasible action sequences. Our results suggest that ASCII rendering can serve as a lightweight, interpretable modality bridge from images to text, complementing conventional VLA pipelines, and opening directions for VLA research with text-only backbones.