Research map / Vision-language and multimodal / Multimodal LLMs
EmbedCopilot: Evaluating Vision-Language Models for Hardware-Aware Embedded System Development
Dongsheng Yuan, Yimo Deng, Huangxun Chen
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add image inputs (ablation) | GPT-4o · EmbedCopilot-Bench | LLMall: 80.3 -> 87.0 | Improved |
| add image inputs (ablation) | GPT-4o · EmbedCopilot-Bench | LLMhw: 65.6 -> 84.7 | Improved |
| add image inputs (ablation) | Qwen2.5-VL-7B · EmbedCopilot-Bench | LLMhw: 43.2 -> 59.0 | Improved |
| add image inputs (ablation) | Qwen2.5-VL-7B · EmbedCopilot-Bench | LLMcode: 56.1 -> 66.1 | Improved |
| add image inputs (ablation) | Qwen3-VL-8B · EmbedCopilot-Bench | LLMhw: 57.4 -> 71.9 | Improved |
| add image inputs (ablation) | Gemma-3-4B · EmbedCopilot-Bench | LLMall: 62.6 -> 62.2 | No clear effect |
| add image inputs (ablation) | LLaVA-OneVision-7B · EmbedCopilot-Bench | LLMall: 61.5 -> 60.6 | No clear effect |
| add image inputs (ablation) | GPT-5 · EmbedCopilot-Bench | Estr: 17.2% -> 31.0% | Improved |
| add image inputs (ablation) | Claude-Sonnet-4.5 · EmbedCopilot-Bench | Estr: 48.3% -> 44.8% | Hurt |
| scale Qwen3-VL from 2B to 4B to 8B with images | Qwen3-VL · EmbedCopilot-Bench | LLMall: 54.9 -> 65.6 -> 74.0 | Improved |
| allow pin-index errors (permissive criterion) | Mistral-Small-3-24B · EmbedCopilot-Bench | execution success: 37.9% -> 82.8% | Improved |
| allow pin-index errors (permissive criterion) | Claude-Sonnet-4.5 · EmbedCopilot-Bench | execution success: 44.8% -> 86.2% | Improved |
Closest papers by meaning
- VLsI: Verbalized Layers-to-Interactions from Large to Small Vision Language Models (CVPR 2025 · similarity 0.86)
- Thinking with Programming Vision: Towards a Unified View for Thinking with Images (CVPR 2026 · similarity 0.85)
- PCEval: A Benchmark for Evaluating Physical Computing Capabilities of Large Language Models (NeurIPS 2026 · similarity 0.84)
- VL-RouterBench: A Benchmark for Vision-Language Model Routing (CVPR 2026 · similarity 0.84)
- VCG-Bench: Towards A Unified Visual-Centric Benchmark for Structured Generation and Editing (ICML 2026 · similarity 0.84)
- DeepRTL: Bridging Verilog Understanding and Generation with a Unified Representation Model (ICLR 2025 · similarity 0.84)
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.
