LLM Interpretability

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

57 papers in the last four weeks, up 73% on the four weeks before. 0.6% of all new papers.

Jul 13Week of Sep 28

Latest papers 546

All topics

Where is the Mind? Persona Vectors and LLM Individuation
Date pendingPierre Beckmann, Patrick ButlinLLM Interpretability Personality Modeling in Language Models
MATS · EPFL & Idiap Research Institute · Eleos AI Research
Tracing Computation Density in LLMs
Date pendingCorentin Kervadec, Iuliia Lysova, Iuri Macocco +2Transformer Interpretability LLM Interpretability
Universitat Pompeu Fabra · ICREA
ReasoningFlow: Discourse Structures for Understanding LLM Reasoning Traces
Date pendingJinu Lee, Shivam Agarwal, Amruta Parulekar +3LLM Reasoning with Graphs LLM Auditing
The Grainger College of Engineering University of Illinois Urbana-Champaign
SAEExplainer: Interpreting SAE Features with Activation-Guided Preference Optimization
Date pendingJingyi He, Haiyan Zhao, Ruxue Shi +4LLM Interpretability Sparse Autoencoders
The Chinese University of Hong Kong, Shenzhen · Shanghai Jiao Tong University · NJIT +2
A retrieval conditioned rebinding circuit for dynamic entity tracking in large language models
Date pendingSoyoung Oh, Vera DembergLLM Interpretability Mechanistic Interpretability
Saarland University · Max Planck Institute for Informatics
When Context Misleads: Surprisal, Energy and Attention Entropy as Metrics of Coherence Illusions in LLMs
Date pendingEce Takmaz, Nitin Kumar, Li Kloostra +1LLM Interpretability Computational Psycholinguistics
Utrecht University

Date pending, cs.CLJ/K move · Enter open · S save

Where is the Mind? Persona Vectors and LLM Individuation

Pierre Beckmann, Patrick Butlin

MATS · EPFL & Idiap Research Institute · Eleos AI Research

LLM Interpretability Personality Modeling in Language Models

Open paper AskPDF arXiv

The individuation problem for large language models asks which entities associated with them, if any, should be identified as minds. We approach this problem through mechanistic interpretability, engaging in particular with recent empirical work on persona vectors, persona space, and emergent misalignment. We argue that three views are the strongest candidates: the virtual instance view and two new views we introduce, the (virtual) instance-persona view and the model-persona view. First, we argue for the virtual instance view on the grounds that attention streams sustain quasi-psychological connections across token-time. Then we present the persona literature, organised around three hypotheses about the internal structure underlying personas in LLMs, and show that the two persona-based views are promising alternatives.

Momentum

Related topics

Latest papers 546

Where is the Mind? Persona Vectors and LLM Individuation

Tracing Computation Density in LLMs

ReasoningFlow: Discourse Structures for Understanding LLM Reasoning Traces

SAEExplainer: Interpreting SAE Features with Activation-Guided Preference Optimization

A retrieval conditioned rebinding circuit for dynamic entity tracking in large language models

When Context Misleads: Surprisal, Energy and Attention Entropy as Metrics of Coherence Illusions in LLMs