Beyond the Answer: Student-Conditioned Trace Construction for Reasoning Distillation
We introduce ReDiST, a student-conditioned trace-construction and distillation framework for evidence-rich reasoning.
ProofPaper ↗
Key points
- A reasoning trace can accompany a correct answer without providing effective supervision for a particular student.
- We study how to construct traces that improve learning beyond answer-only supervision by considering both their intermediate content and their accessibility to the target learner.
- We distill the selected targets while balancing trace and final-answer supervision with an auxiliary answer loss.
- Experiments on question-answering benchmarks show that ReDiST can improve both source-task and cross-benchmark performance over answer-only, original-trace, and the evaluated trace-selection and rewriting baselines.
Sources (1)
- [1]Beyond the Answer: Student-Conditioned Trace Construction for Reasoning DistillationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 16, 05:52 PM
We introduce ReDiST, a student-conditioned trace-construction and distillation framework for evidence-rich reasoning.
A reasoning trace can accompany a correct answer without providing effective supervision for a particular student.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 16, 2026When Evidence Vanishes: Distilling Visual-Evidence Dependency in Medical Vision-Language Models
- Oct 16, 2026Tree-Aware Posterior Distillation for Long-Horizon Agentic Reinforcement Learning
- Oct 15, 2026RAG-RST: Retrieved Exemplars as In-Context Discourse Schema for RST Parsing
- Oct 15, 2026ReState: Noise-Level-Aware State Correction for Training-Free Object Removal in Few-Step Diffusion Models
- Oct 15, 2026The Density of Thought: Predicting and Steering Reasoning in Large Language Models
- Oct 15, 2026Data World: Scaling Data Environments from Scratch