AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Hints of Prompt: Enhancing Visual Representation for Multimodal LLMs in Autonomous Driving

Hao Zhou, Zhanning Gao, Zhili Chen, Maosheng Ye, Qifeng Chen, Tongyi Cao, Honggang Qi

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Instruction-tuned multimodal assistants · 3 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
add Hints of Prompt (three hints) to LLaVA-v1.5LLaVA-v1.5 · LingoQALingo-Judge: 63.2 -> 67.8Improved
add HoP to LLaVA-v1.5LLaVA-v1.5 · BDD-X (All)CIDEr: 96.51 -> 102.20Improved
add HoP to LLaVA-v1.5LLaVA-v1.5 · DRAMABLEU-4: 56.1 -> 56.3Improved
train HoP with only 25% data (ablation)HoP · LingoQALingo-Judge: 64.0 vs full-data LLaVA-v1.5 63.2Improved
use Joint Cross-Attention instead of concatenation to fuse hints (ablation)HoP · LingoQALingo-Judge: 58.8 -> 67.8Improved
use Joint Cross-Attention instead of Self-Cross-Attention (ablation)HoP · LingoQALingo-Judge: 65.2 -> 67.8Improved
use Joint Cross-Attention instead of Sequential Cross-Attention (ablation)HoP · LingoQALingo-Judge: 67.0 -> 67.8Improved
use Mask2Former instead of GroundingDINO for Semantic hint (ablation)HoP · LingoQALingo-Judge: 66.4 -> 67.8Improved
take Question hint from LLM embeddings instead of CLIP text (ablation)HoP · LingoQALingo-Judge: 65.2 -> 67.8Improved
increase Semantic hint tokens from 8 to 16 (ablation)HoP · LingoQAall metricsImproved
increase Semantic hint tokens from 32 to 64 (ablation)HoP · LingoQAall metricsHurt
use DINOv2 similarity tokens instead of full DINOv2 tokens for Affinity hint (ablation)HoP · LingoQACIDEr: 66.0 -> 71.1Mixed

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