Crimson AI NewsA CrimsonLingua Network service
EN ع
← Back to news
Research paper Hugging Face

TTPO: Label-Free Test-Time Training Matches Supervised Math Reasoning

AI By Crimson AI Hugging Face Papers 28 August 2026 · 00:00 2 views
Share: X Telegram

Hugging Face researchers introduce Test-Time Policy Optimization (TTPO), a label-free method that distills agreeing rollouts and penalizes disagreeing ones, matching supervised performance on five competition-level benchmarks.

TTPO: Label-Free Test-Time Training Matches Supervised Math Reasoning

Key points

Recent advances in mathematical reasoning for large language models have been driven by post-training methods like Reinforcement Learning (RL) and On-Policy Self-Distillation (OPSD). However, these methods typically rely on ground-truth labels, which prevents their use in test-time training (TTT) scenarios where labels are unavailable.

A natural alternative is to replace ground truth with majority-vote pseudo-labels, but this approach is fragile: an incorrect vote can corrupt the teacher and mislead every token. The researchers observed that this failure mode is asymmetric—rollouts that disagree with the pseudo-label are usually wrong, regardless of whether the vote itself is correct.

Building on this insight, they propose Test-Time Policy Optimization (TTPO), an asymmetric objective that distills agreeing rollouts via OPSD and penalizes disagreeing rollouts with Grouped RL. Token-level selection further refines both branches: distillation down-weights already-converged positions, while RL penalizes only confident errors.

Without any labels, TTPO matches label-supervised OPSD on five competition-level benchmarks. It raises Qwen3-1.7B from 38.0% to 45.2% in TTT, yields +25.2% to +36.4% without thinking, and shows strong cross-task generalization.

BenchmarkSupervised OPSDTTPO (label-free)
Competition-level benchmarks (5)BaselineMatches supervised
Qwen3-1.7B TTT accuracy38.0%45.2%
Without thinking+25.2% to +36.4%
Source
Hugging Face · Hugging Face Papers
Related news
Research paper
Hugging Face 29 Aug 2026

Hugging Face Audit: 110 of 124 AI Evaluations Fail to Support Their Claims

A new commit-bound census of 124 Inspect Evals units reveals that 110 stop before deterministic inference due to missing historica...

4
Research paper
Hugging Face 29 Aug 2026

Aphanta: New Framework Diagnoses When Image Editing Boosts Multimodal Reasoning

Hugging Face researchers introduce Aphanta, a diagnostic framework that evaluates when image-editing intermediates improve multimo...

5
Research paper
Hugging Face 29 Aug 2026

EditaLive! Enables Real-Time Character Video Editing for Live Streaming

Hugging Face researchers introduce EditaLive, a framework for real-time human-centric video editing in live streams, achieving sta...

4