Imagen de Portada

Thought Branches: Interpreting LLM Reasoning Requires Resampling

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Macar, Uzay, Bogdan, Paul C., Rajamanoharan, Senthooran, Nanda, Neel
Formato:	Preprint
Publicado:	2025
Materias:	Machine Learning Artificial Intelligence Computation and Language
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Thought Anchors: Which LLM Reasoning Steps Matter?
por: Bogdan, Paul C., et al.
Publicado: (2025)

Chain-of-Thought Reasoning In The Wild Is Not Always Faithful
por: Arcuschin, Iván, et al.
Publicado: (2025)

Do I Know This Entity? Knowledge Awareness and Hallucinations in Language Models
por: Ferrando, Javier, et al.
Publicado: (2024)

Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning
por: Casademunt, Helena, et al.
Publicado: (2025)

Emergent Misalignment is Easy, Narrow Misalignment is Hard
por: Soligo, Anna, et al.
Publicado: (2026)

Convergent Linear Representations of Emergent Misalignment
por: Soligo, Anna, et al.
Publicado: (2025)

Simple Mechanistic Explanations for Out-Of-Context Reasoning
por: Wang, Atticus, et al.
Publicado: (2025)

Model Organisms for Emergent Misalignment
por: Turner, Edward, et al.
Publicado: (2025)

Are Sparse Autoencoders Useful? A Case Study in Sparse Probing
por: Kantamneni, Subhash, et al.
Publicado: (2025)

Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2
por: Lieberum, Tom, et al.
Publicado: (2024)

How Well Do Models Follow Their Constitutions?
por: Jakkli, Arya, et al.
Publicado: (2026)

Dense SAE Latents Are Features, Not Bugs
por: Sun, Xiaoqing, et al.
Publicado: (2025)

Explorations of Self-Repair in Language Models
por: Rushing, Cody, et al.
Publicado: (2024)

Towards Best Practices of Activation Patching in Language Models: Metrics and Methods
por: Zhang, Fred, et al.
Publicado: (2023)

Overcoming Sparsity Artifacts in Crosscoders to Interpret Chat-Tuning
por: Minder, Julian, et al.
Publicado: (2025)

Improving Dictionary Learning with Gated Sparse Autoencoders
por: Rajamanoharan, Senthooran, et al.
Publicado: (2024)

Hierarchical Multi-Armed Bandits for the Concurrent Intelligent Tutoring of Concepts and Problems of Varying Difficulty Levels
por: Castleman, Blake, et al.
Publicado: (2024)

Towards eliciting latent knowledge from LLMs with mechanistic interpretability
por: Cywiński, Bartosz, et al.
Publicado: (2025)

ReasonFlux: Hierarchical LLM Reasoning via Scaling Thought Templates
por: Yang, Ling, et al.
Publicado: (2025)

Weight-of-Thought Reasoning: Exploring Neural Network Weights for Enhanced LLM Reasoning
por: Punjwani, Saif, et al.
Publicado: (2025)

Transcoders Find Interpretable LLM Feature Circuits
por: Dunefsky, Jacob, et al.
Publicado: (2024)

Scaling sparse feature circuit finding for in-context learning
por: Kharlapenko, Dmitrii, et al.
Publicado: (2025)

What's the plan? Metrics for implicit planning in LLMs and their application to rhyme generation and question answering
por: Maar, Jim, et al.
Publicado: (2026)

Deep Dense Exploration for LLM Reinforcement Learning via Pivot-Driven Resampling
por: Guo, Yiran, et al.
Publicado: (2026)

Sketch-of-Thought: Efficient LLM Reasoning with Adaptive Cognitive-Inspired Sketching
por: Aytes, Simon A., et al.
Publicado: (2025)

Subliminal Learning Is Steering Vector Distillation
por: Blank, Camila, et al.
Publicado: (2026)

ORION: Teaching Language Models to Reason Efficiently in the Language of Thought
por: Tanmay, Kumar, et al.
Publicado: (2025)

A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoning
por: Hu, Lijie, et al.
Publicado: (2024)

Real-Time Detection of Hallucinated Entities in Long-Form Generation
por: Obeso, Oscar, et al.
Publicado: (2025)

Censored LLMs as a Natural Testbed for Secret Knowledge Elicitation
por: Casademunt, Helena, et al.
Publicado: (2026)

Diffusion of Thoughts: Chain-of-Thought Reasoning in Diffusion Language Models
por: Ye, Jiacheng, et al.
Publicado: (2024)

Fractured Chain-of-Thought Reasoning
por: Liao, Baohao, et al.
Publicado: (2025)

Interpreting and Controlling LLM Reasoning through Integrated Policy Gradient
por: Li, Changming, et al.
Publicado: (2026)

When Chain of Thought is Necessary, Language Models Struggle to Evade Monitors
por: Emmons, Scott, et al.
Publicado: (2025)

SmartSwitch: Advancing LLM Reasoning by Overcoming Underthinking via Promoting Deeper Thought Exploration
por: Zhang, Xichen, et al.
Publicado: (2025)

Reject, Resample, Repeat: Understanding Parallel Reasoning in Language Model Inference
por: Golowich, Noah, et al.
Publicado: (2026)

Reasoning to Learn from Latent Thoughts
por: Ruan, Yangjun, et al.
Publicado: (2025)

Confidence Regulation Neurons in Language Models
por: Stolfo, Alessandro, et al.
Publicado: (2024)

Building Production-Ready Probes For Gemini
por: Kramár, János, et al.
Publicado: (2026)

Refusal in Language Models Is Mediated by a Single Direction
por: Arditi, Andy, et al.
Publicado: (2024)