AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Boosting MLLM Reasoning with Text-Debiased Hint-GRPO

Qihan Huang, Weilong Dai, Jinlong Liu, Wanggui He, Hao Jiang, Mingli Song, Jingyuan Chen, Chang Yao, Jie Song

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Chain-of-thought multimodal reasoning · 21 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
convert multi-choice questions to fill-in-the-blank in training data (ablation)GRPO on Qwen2-VL-7B · Geo170Kaccuracy (%): 35.81 -> 38.46Improved
inject hint in the answer rather than in the query (ablation)Hint-GRPO on Qwen2-VL-7B · Geo170Kaccuracy (%): 41.64 -> 45.62Improved
add adaptive hints (Hint-GRPO) to GRPOQwen2-VL-7B · Geo170Kaccuracy (%): 38.46 -> 45.62Improved
add adaptive hints (Hint-GRPO) to GRPOQwen2.5-VL-3B · Geo170Kaccuracy (%): 45.49 -> 53.32Improved
add text-bias calibration at test time (ablation)Hint-GRPO on Qwen2-VL-7B · Geo170Kaccuracy (%): 45.62 -> 46.68Improved
add text-bias calibration at test time (ablation)Hint-GRPO on Qwen2.5-VL-3B · Geo170Kaccuracy (%): 53.32 -> 55.31Improved
raise calibration strength gamma from 0.8 to 1.6 (ablation)Debiased Hint-GRPO on Qwen2-VL-7B · Geo170Kaccuracy (%): 46.68 -> 44.69Hurt
increase number of hint groups M from 1 to 3 (ablation)Hint-GRPO on Qwen2-VL-7B · Geo170Kaccuracy (%): 38.46 -> 45.62Improved
increase number of hint groups M from 3 to 4 (ablation)Hint-GRPO on Qwen2-VL-7B · Geo170Kaccuracy (%): 45.62 -> 45.89No clear effect
increase number of hint groups M from 3 to 4 (ablation)Hint-GRPO on Qwen2.5-VL-3B · Geo170Kaccuracy (%): 53.32 -> 52.65Hurt
use adaptive M=3 hint groups in place of M=1 (no hints) (ablation)Hint-GRPO on Qwen2.5-VL-3B · Geo170Kaccuracy (%): 45.49 -> 53.32Improved
use SFT instead of the original modelQwen2-VL-7B · Geometry average (7 datasets)accuracy (%): 30.40 -> 25.50Hurt

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