Cover Image

How and Why LLMs Generalize: A Fine-Grained Analysis of LLM Reasoning from Cognitive Behaviors to Low-Level Patterns

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Bai, Haoyue, Sun, Yiyou, Hu, Wenjie, Qiu, Shi, Huan, Maggie Ziyu, Song, Peiyang, Nowak, Robert, Song, Dawn
Format:	Preprint
Published:	2025
Subjects:	Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Climbing the Ladder of Reasoning: What LLMs Can-and Still Can't-Solve after SFT?
by: Sun, Yiyou, et al.
Published: (2025)

RL Grokking Recipe: How Does RL Unlock and Transfer New Algorithms in LLMs?
by: Sun, Yiyou, et al.
Published: (2025)

Why and How LLMs Hallucinate: Connecting the Dots with Subsequence Associations
by: Sun, Yiyou, et al.
Published: (2025)

The Long-Horizon Task Mirage? Diagnosing Where and Why Agentic Systems Break
by: Wang, Xinyu Jessica, et al.
Published: (2026)

Can Aha Moments Be Fake? Towards Quantifying Decorative and True Thinking in Chain-of-Thought
by: Zhao, Jiachen, et al.
Published: (2025)

OMEGA: Can LLMs Reason Outside the Box in Math? Evaluating Exploratory, Compositional, and Transformative Generalization
by: Sun, Yiyou, et al.
Published: (2025)

Where's the liability in the Generative Era? Recovery-based Black-Box Detection of AI-Generated Content
by: Bai, Haoyue, et al.
Published: (2025)

Strategy Executability in Mathematical Reasoning: Leveraging Human-Model Differences for Effective Guidance
by: Liang, Weida, et al.
Published: (2026)

MIRAGE-Bench: LLM Agent is Hallucinating and Where to Find Them
by: Zhang, Weichen, et al.
Published: (2025)

Can LLMs Ask Good Questions?
by: Zhang, Yueheng, et al.
Published: (2025)

ReasonMap: Towards Fine-Grained Visual Reasoning from Transit Maps
by: Feng, Sicheng, et al.
Published: (2025)

AHA: Human-Assisted Out-of-Distribution Generalization and Detection
by: Bai, Haoyue, et al.
Published: (2024)

Large Language Model Reasoning Failures
by: Song, Peiyang, et al.
Published: (2026)

LazyEviction: Lagged KV Eviction with Attention Pattern Observation for Efficient Long Reasoning
by: Zhang, Haoyue, et al.
Published: (2025)

Unsafer in Many Turns: Benchmarking and Defending Multi-Turn Safety Risks in Tool-Using Agents
by: Li, Xu, et al.
Published: (2026)

The Personality Illusion: Revealing Dissociation Between Self-Reports & Behavior in LLMs
by: Han, Pengrui, et al.
Published: (2025)

Deep Active Learning in the Open World
by: Xie, Tian, et al.
Published: (2024)

FineCog-Nav: Integrating Fine-grained Cognitive Modules for Zero-shot Multimodal UAV Navigation
by: Shao, Dian, et al.
Published: (2026)

Integrating Fine-Grained Audio-Visual Evidence for Robust Multimodal Emotion Reasoning
by: Zhao, Zhixian, et al.
Published: (2026)

Are You Getting What You Pay For? Auditing Model Substitution in LLM APIs
by: Cai, Will, et al.
Published: (2025)

Towards Robust Out-of-Distribution Generalization: Data Augmentation and Neural Architecture Search Approaches
by: Bai, Haoyue
Published: (2024)

In-Context Learning May Not Elicit Trustworthy Reasoning: A-Not-B Errors in Pretrained Language Models
by: Han, Pengrui, et al.
Published: (2024)

When and How Does In-Distribution Label Help Out-of-Distribution Detection?
by: Du, Xuefeng, et al.
Published: (2024)

On Fine-Grained I/O Complexity of Attention Backward Passes
by: Li, Xiaoyu, et al.
Published: (2024)

Securing LLM Agents Need Intent-to-Execution Integrity
by: Qu, Wenjie, et al.
Published: (2026)

SSR3D-LLM: Structured Spatial Reasoning via Latent Steps for Fine-Grained Grounding in Unified 3D-LLMs
by: Li, Jiawei, et al.
Published: (2026)

Fine-Grained Self-Endorsement Improves Factuality and Reasoning
by: Wang, Ante, et al.
Published: (2024)

HoneyComb: A Flexible LLM-Based Agent System for Materials Science
by: Zhang, Huan, et al.
Published: (2024)

Fine-Tune an SLM or Prompt an LLM? The Case of Generating Low-Code Workflows
by: Ayala, Orlando Marquez, et al.
Published: (2025)

CTIArena: Benchmarking LLM Knowledge and Reasoning Across Heterogeneous Cyber Threat Intelligence
by: Cheng, Yutong, et al.
Published: (2025)

The Alignment Floor: How Persona Customization Breaks Safety in Weakly-Aligned LLMs
by: Zhang, Xing, et al.
Published: (2026)

Learning to Verify Summary Facts with Fine-Grained LLM Feedback
by: Oh, Jihwan, et al.
Published: (2024)

HeartLLM: Discretized ECG Tokenization for LLM-Based Diagnostic Reasoning
by: Yang, Jinning, et al.
Published: (2025)

Fine-Grained Instruction-Guided Graph Reasoning for Vision-and-Language Navigation
by: Liu, Yaohua, et al.
Published: (2025)

Self-Sovereign Agent
by: Qu, Wenjie, et al.
Published: (2026)

Patterns Over Principles: The Fragility of Inductive Reasoning in LLMs under Noisy Observations
by: Li, Chunyang, et al.
Published: (2025)

Measuring Reasoning Quality in LLMs: A Multi-Dimensional Behavioral Framework
by: Şenol, Ali, et al.
Published: (2026)

Improving LLM Safety Alignment with Dual-Objective Optimization
by: Zhao, Xuandong, et al.
Published: (2025)

Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM Reasoning
by: Huan, Maggie, et al.
Published: (2025)

Cross-Level Multi-Instance Distillation for Self-Supervised Fine-Grained Visual Categorization
by: Bi, Qi, et al.
Published: (2024)