Cover Image

Generalization or Hallucination? Understanding Out-of-Context Reasoning in Transformers

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Huang, Yixiao, Zhu, Hanlin, Guo, Tianyu, Jiao, Jiantao, Sojoudi, Somayeh, Jordan, Michael I., Russell, Stuart, Mei, Song
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

How Do LLMs Perform Two-Hop Reasoning in Context?
by: Guo, Tianyu, et al.
Published: (2025)

Transformers Provably Learn to Internalize Chain-of-Thought
by: Huang, Yixiao, et al.
Published: (2026)

Towards a Theoretical Understanding of the 'Reversal Curse' via Training Dynamics
by: Zhu, Hanlin, et al.
Published: (2024)

GSM-Agent: Understanding Agentic Reasoning Using Controllable Environments
by: Zhu, Hanlin, et al.
Published: (2025)

Efficient Prompt Caching via Embedding Similarity
by: Zhu, Hanlin, et al.
Published: (2024)

Iterative Data Smoothing: Mitigating Reward Overfitting and Overoptimization in RLHF
by: Zhu, Banghua, et al.
Published: (2024)

Reinforcement Learning for Flow-Matching Policies
by: Pfrommer, Samuel, et al.
Published: (2025)

Towards Optimal Statistical Watermarking
by: Huang, Baihe, et al.
Published: (2023)

Reasoning by Superposition: A Theoretical Perspective on Chain of Continuous Thought
by: Zhu, Hanlin, et al.
Published: (2025)

Ranking Manipulation for Conversational Search Engines
by: Pfrommer, Samuel, et al.
Published: (2024)

Token Assorted: Mixing Latent and Text Tokens for Improved Language Model Reasoning
by: Su, DiJia, et al.
Published: (2025)

Spooky Action at a Distance: Normalization Layers Enable Side-Channel Spatial Communication
by: Pfrommer, Samuel, et al.
Published: (2025)

Multi-Objective Learning for Diffusion Models: A Statistical Theory under Semi-Supervised Learning
by: Cheng, Ziheng, et al.
Published: (2026)

Active-Dormant Attention Heads: Mechanistically Demystifying Extreme-Token Phenomena in LLMs
by: Guo, Tianyu, et al.
Published: (2024)

OVERT: A Benchmark for Over-Refusal Evaluation on Text-to-Image Models
by: Cheng, Ziheng, et al.
Published: (2025)

Out-of-Context Reasoning in Large Language Models
by: Shaki, Jonathan, et al.
Published: (2025)

Emergence of Superposition: Unveiling the Training Dynamics of Chain of Continuous Thought
by: Zhu, Hanlin, et al.
Published: (2025)

Large Language Models are Limited in Out-of-Context Knowledge Reasoning
by: Hu, Peng, et al.
Published: (2024)

Generative AI Security: Challenges and Countermeasures
by: Zhu, Banghua, et al.
Published: (2024)

HaploVL: A Single-Transformer Baseline for Multi-Modal Understanding
by: Yang, Rui, et al.
Published: (2025)

Vision-and-Language Navigation Generative Pretrained Transformer
by: Hanlin, Wen
Published: (2024)

Facilitating Long Context Understanding via Supervised Chain-of-Thought Reasoning
by: Lin, Jingyang, et al.
Published: (2025)

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

Principled Reinforcement Learning with Human Feedback from Pairwise or $K$-wise Comparisons
by: Zhu, Banghua, et al.
Published: (2023)

Hallucination Detection and Hallucination Mitigation: An Investigation
by: Luo, Junliang, et al.
Published: (2024)

On Representation Complexity of Model-based and Model-free Reinforcement Learning
by: Zhu, Hanlin, et al.
Published: (2023)

Smoothing Out Hallucinations: Mitigating LLM Hallucination with Smoothed Knowledge Distillation
by: Nguyen, Hieu, et al.
Published: (2025)

Toward a Theory of Tokenization in LLMs
by: Rajaraman, Nived, et al.
Published: (2024)

Credal Transformer: A Principled Approach for Quantifying and Mitigating Hallucinations in Large Language Models
by: Ji, Shihao, et al.
Published: (2025)

Theoretical Understanding of In-Context Learning in Shallow Transformers with Unstructured Data
by: Xing, Yue, et al.
Published: (2024)

Rethinking Visual Neglect: Steering via Context-Preference for MLLM Hallucination Mitigation
by: Wu, Jingwen, et al.
Published: (2026)

Thinking LLMs: General Instruction Following with Thought Generation
by: Wu, Tianhao, et al.
Published: (2024)

Understanding Contextual Recall in Transformers: How Finetuning Enables In-Context Reasoning over Pretraining Knowledge
by: Vasudeva, Bhavya, et al.
Published: (2026)

Evaluating Multilingual Long-Context Models for Retrieval and Reasoning
by: Agrawal, Ameeta, et al.
Published: (2024)

Extractive Structures Learned in Pretraining Enable Generalization on Finetuned Facts
by: Feng, Jiahai, et al.
Published: (2024)

Transformers as Decision Makers: Provable In-Context Reinforcement Learning via Supervised Pretraining
by: Lin, Licong, et al.
Published: (2023)

RAGulator: Lightweight Out-of-Context Detectors for Grounded Text Generation
by: Poey, Ian, et al.
Published: (2024)

Understanding In-Context Learning Beyond Transformers: An Investigation of State Space and Hybrid Architectures
by: Wang, Shenran, et al.
Published: (2025)

IDA-Bench: Evaluating LLMs on Interactive Guided Data Analysis
by: Li, Hanyu, et al.
Published: (2025)

Developing a Reliable, Fast, General-Purpose Hallucination Detection and Mitigation Service
by: Wang, Song, et al.
Published: (2024)