AI Paper Atlas

Research map / Large language models / Agents, code and math

TASE: Token Awareness and Structured Evaluation for Multilingual Language Models

Chenzhuo Zhao, Xinda Wang, Yue Huang, Junting Lu, Ziqian Liu

AAAI 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
fine-tune with GRPO on synthetic TASE data (fine-grained reward)Qwen2.5-14B-Instruct · TASE averageaccuracy (%): 11.72 -> 20.40Improved
use fine-grained reward instead of coarse binary reward in GRPO (ablation)qwen2.5-14b-grpo · TASE averageaccuracy (%): 16.08 -> 20.40Improved
GRPO with coarse binary reward (ablation)Qwen2.5-14B-Instruct · TASE averageaccuracy (%): 11.72 -> 16.08Improved
GRPO fine-tuningQwen2.5-14B-Instruct · TASE LENOPaccuracy (%): 4.87 -> 27.22Improved
add chain-of-thought promptingo1-mini · TASE averageaccuracy (%): 24.56 -> 37.27Improved
add chain-of-thought promptinggpt-3.5-turbo · TASE averageaccuracy (%): 12.80 -> 18.45Improved
add chain-of-thought promptingdoubao-pro · TASE averageaccuracy (%): 19.87 -> 22.94Improved
add chain-of-thought promptingdoubao-lite · TASE averageaccuracy (%): 16.31 -> 16.44No clear effect
add chain-of-thought promptingqwen2.5-14B · TASE averageaccuracy (%): 11.72 -> 13.69Improved
evaluate in Chinese or Korean instead of Englisho3 · TASE language splitaccuracy (%): 86.71 -> 69.12 (Chinese), 67.83 (Korean)Hurt
evaluate in Chinese or Korean instead of Englishgpt-4.1 · TASE language splitaccuracy (%): 39.89 -> 27.87 (Chinese), 25.75 (Korean)Hurt
increase model size within Yi-1.5, Qwen2.5 and Qwen3 familiesYi-1.5, Qwen2.5, Qwen3 · TASE averageaccuracy (%)Mixed

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