Failure Profiling and Reachable Trajectory Selection for Reasoning Distillation
cs.CL, cs.AI, cs.LG
Submitted: 2026-11-06
Updated: 2026-11-06
Code: https://github.com/visresearch/FPRT
Terminology
Sources
- GPT-4 Technical Report
- Learning From Mistakes Makes LLM Better Reasoner
- Training Verifiers to Solve Math Word Problems
- Beyond Rejection Sampling: Trajectory Fusion for Scaling Mathematical Reasoning
- The Llama 3 Herd of Models
- rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- UNDO: Understanding Distillation as Optimization
- The Signal is in the Steps: Local Scoring for Reasoning Data Selection
- Common 7B Language Models Already Possess Strong Math Capabilities
- DotaMath: Decomposition of Thought with Code Assistance and Self-correction for Mathematical Reasoning
- Step-Controlled DPO: Leveraging Stepwise Error for Enhanced Mathematical Reasoning
- WizardMath: Empowering Mathematical Reasoning for Large Language Models via Reinforced Evol-Instruct
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Can LLMs Learn from Previous Mistakes? Investigating LLMs' Errors to Boost for Reasoning
- Enhancing LLM Reasoning via Critique Models with Test-Time and Training-Time Supervision
- Monte Carlo Tree Search Boosts Reasoning via Iterative Preference Learning
- Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement
- Scaling Relationship on Learning Mathematical Reasoning with Large Language Models
- LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data