How to Fine-Tune a Reasoning Model? A Teacher-Student Cooperation Framework to Synthesize Student-Consistent SFT Data

Zixian Huang, Kaichen Yang, Xu Huang, Feiyang Hao, Qiming Ge, Bowen Li, He Du, Kai Chen, Qipeng Guo · Mar 23, 2026 · Citations: 0

Open arXiv Find Implementation RSS feed Shortlist (0)

How to use this page

Provisional trust

This page is a lightweight research summary built from the abstract and metadata while deeper extraction catches up.

Best use

Background context only

What to verify

Read the full paper before copying any benchmark, metric, or protocol choices.

Evidence quality

Provisional

Derived from abstract and metadata only.

Abstract

A widely adopted strategy for model enhancement is to use synthetic data generated by a stronger model for supervised fine-tuning (SFT). However, for emerging reasoning models like Qwen3-8B, this approach often fails to improve reasoning capabilities and can even lead to a substantial drop in performance. In this work, we identify substantial stylistic divergence between teacher generated data and the distribution of student as a major factor impacting SFT. To bridge this gap, we propose a Teacher-Student Cooperation Data Synthesis framework (TESSY), which interleaves teacher and student models to alternately generate style and non-style tokens. Consequently, TESSY produces synthetic sequences that inherit the advanced reasoning capabilities of the teacher while maintaining stylistic consistency with the distribution of the student. In experiments on code generation using GPT-OSS-120B as the teacher, fine-tuning Qwen3-8B on teacher-generated data leads to performance drops of 3.25% on LiveCodeBench-Pro and 10.02% on OJBench, whereas TESSY achieves improvements of 11.25% and 6.68%.

Abstract-only analysis — low confidence

All signals on this page are inferred from the abstract only and may be inaccurate. Do not use this page as a primary protocol reference.

This page is still relying on abstract and metadata signals, not a fuller protocol read.

Human Eval Hub LLM-as-Judge Hub Pairwise Preference Hub

Should You Rely On This Paper?

Signal extraction is still processing. This page currently shows metadata-first guidance until structured protocol fields are ready.

Best use

Background context only

Use if you need

A provisional background reference while structured extraction finishes.

Main weakness

This page is still relying on abstract and metadata signals, not a fuller protocol read.

Trust level

Provisional

Usefulness score

Unavailable

Eval-fit score is unavailable until extraction completes.

Human Feedback Signal

Not explicit in abstract metadata

Evaluation Signal

Weak / implicit signal

Usefulness for eval research

Provisional (processing)

Extraction confidence 0%

If you are doing eval pipeline work, start here:

Human Eval Hub LLM-as-Judge Hub Pairwise Preference Hub Tool-Use Eval Hub

What We Could Verify

These are the protocol signals we could actually recover from the available paper metadata. Use them to decide whether this paper is worth deeper reading.

Human Feedback Types

provisional (inferred)

None explicit

No explicit feedback protocol extracted.

"A widely adopted strategy for model enhancement is to use synthetic data generated by a stronger model for supervised fine-tuning (SFT)."

Evaluation Modes

provisional (inferred)

None explicit

Validate eval design from full paper text.

"A widely adopted strategy for model enhancement is to use synthetic data generated by a stronger model for supervised fine-tuning (SFT)."

Quality Controls

provisional (inferred)

Not reported

No explicit QC controls found.

"A widely adopted strategy for model enhancement is to use synthetic data generated by a stronger model for supervised fine-tuning (SFT)."

Benchmarks / Datasets

provisional (inferred)

LiveCodeBench

Useful for quick benchmark comparison.

"In experiments on code generation using GPT-OSS-120B as the teacher, fine-tuning Qwen3-8B on teacher-generated data leads to performance drops of 3.25% on LiveCodeBench-Pro and 10.02% on OJBench, whereas TESSY achieves improvements of 11.25% and 6.68%."

Reported Metrics

provisional (inferred)

Not extracted

No metric anchors detected.

"A widely adopted strategy for model enhancement is to use synthetic data generated by a stronger model for supervised fine-tuning (SFT)."

Rater Population

provisional (inferred)

Unknown

Rater source not explicitly reported.

"A widely adopted strategy for model enhancement is to use synthetic data generated by a stronger model for supervised fine-tuning (SFT)."

Human Feedback Details

This page is using abstract-level cues only right now. Treat the signals below as provisional.

Potential human-data signal: No explicit human-data keywords detected.
Potential benchmark anchors: LiveCodeBench
Abstract highlights: 3 key sentence(s) extracted below.

Evaluation Details

Evaluation fields are inferred from the abstract only.

Potential evaluation modes: No explicit eval keywords detected.
Potential metric signals: No metric keywords detected.
Confidence: Provisional (metadata-only fallback).

Research Brief

Metadata summary

A widely adopted strategy for model enhancement is to use synthetic data generated by a stronger model for supervised fine-tuning (SFT).

Based on abstract + metadata only. Check the source paper before making high-confidence protocol decisions.

Key Takeaways

A widely adopted strategy for model enhancement is to use synthetic data generated by a stronger model for supervised fine-tuning (SFT).
However, for emerging reasoning models like Qwen3-8B, this approach often fails to improve reasoning capabilities and can even lead to a substantial drop in performance.
In this work, we identify substantial stylistic divergence between teacher generated data and the distribution of student as a major factor impacting SFT.

Researcher Actions

Compare this paper against others mentioning LiveCodeBench.
Check the full text for explicit evaluation design choices (raters, protocol, and metrics).
Use related-paper links to find stronger protocol-specific references.

Caveats

Generated from abstract + metadata only; no PDF parsing.
Signals below are heuristic and may miss details reported outside the abstract.

Recommended Queries

LiveCodeBench

Related Papers

Papers are ranked by protocol overlap, extraction signal alignment, and semantic proximity.

No related papers found for this item yet.

Get Started

Join the #1 Platform for AI Training Talent

Where top AI builders and expert AI Trainers connect to build the future of AI.

Self-Service

Post a Job

Post your project and get a shortlist of qualified AI Trainers and Data Labelers. Hire and manage your team in the tools you already use.

Create Account & Post a Job

Managed Service

For Large Projects

Done-for-You

We recruit, onboard, and manage a dedicated team inside your tools. End-to-end operations for large or complex projects.

Learn About Managed Service

For Freelancers

Join as an AI Trainer

Find AI training and data labeling projects across platforms, all in one place. One profile, one application process, more opportunities.

Join Now