AI Paper Atlas

Research map / Large language models / Reasoning and chain of thought

Time-Frequency Token Advantage Clipping for Training Efficient Large Reasoning Model

Rong Bao, Bo Wang, Xiao Wang, Hongyu Li, Rui Zheng, Leszek Rutkowski, Qi Zhang, Liang Ding, Dacheng Tao

AAAI 2026 · Large language models › Reasoning and chain of thought · approach: Chain-of-thought training and analysis · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes time-frequency token advantage clipping to train large reasoning models with token-level control over long chain-of-thought, improving reasoning efficiency.

Problem

Long chain-of-thought reasoning in large reasoning models is inefficient: models overthink simple problems or underthink complex ones, and sequence-level length penalties are too coarse to separate core logic from verbose language.

Approach

Applies token-level advantage clipping guided by time-frequency analysis of tokens so that core reasoning tokens are distinguished from verbose language, instead of sequence-level length penalties.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
remove historical token count (time dimension) (ablation)TFAC Qwen3-8B · AIME (Table 4)Avg@32: 71.4 -> 73.9Improved
add intra-path frequency clipping (ablation)TFAC Qwen3-8B · AIME (Table 4)Avg@32: 70.8 -> 73.9Improved
add time-count budget (ablation)TFAC R1-Distill-32B · AIME (Table 4)Avg@32: 68.4 -> 71.2Improved
add frequency clipping (ablation)TFAC R1-Distill-32B · AIME (Table 4)Avg@32: 68.9 -> 71.2Improved
add dynamic sampling (ablation)TFAC · AIME (Table 4)Avg@32: 73.9 -> 74.2 (Qwen3-8B); 71.2 -> 72.9 (R1-Distill-32B)Improved
apply penalty to advantage clipping instead of terminal reward (TFRS) (ablation)TFAC · AIME 2024/2025 (Table 2)Avg@32: 71.9 -> 73.9 (Qwen3-8B, gamma/lambda 0.8); 62.2 -> 66.0 (R1-Distill-32B)Improved
TFAC vs DAPO w. forking tokenTFAC · Table 1 benchmarks, Qwen3-8BAvg@32: MATH-500 95.32 -> 94.96; AIME 2024 77.60 -> 77.80; LCB 58.52 -> 58.15Mixed
TFAC vs DAPO w. forking tokenTFAC · Table 1 benchmarks, R1-Distill-32BAvg@32: MATH-500 96.90 -> 97.13; AIME 2025 53.20 -> 55.50Improved
TFAC vs explicit length-penalty methodsTFAC · Table 1Avg@32: Qwen3-8B AIME 2025: HAPO 58.00 -> TFAC 69.50Improved
TFAC vs untrained base on Qwen3-8B MATH-500TFAC · MATH-500, Qwen3-8BAvg@32: 94.25 -> 94.96No clear effect
TFAC vs DAPO w. forkTFAC · GPQA (OOD)Avg@32: 64.92 -> 66.65 (Qwen3-8B); 63.52 -> 65.35 (R1-Distill-32B)Improved
TFAC under small generation budget (~8k)TFAC · MATH reasoning, token budget scalingperformanceHurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy