Research map / Large language models / Instruction and fine-tuning
GPT4Tools: Teaching Large Language Model to Use Tools via Self-instruction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| fine-tune with GPT4Tools data via LoRA | Vicuna-13B · Seen tools | SR: 12.4 -> 94.1 | Helped |
| fine-tune with GPT4Tools data via LoRA | OPT-13B · Seen tools | SR: 0.0 -> 93.2 | Helped |
| fine-tune with GPT4Tools data via LoRA | LLaMA-13B · Seen tools | SR: 3.2 -> 66.4 | Helped |
| fine-tune with GPT4Tools (zero-shot on unseen tools) | Vicuna-13B · Unseen tools | SR: 26.2 -> 90.6 | Helped |
| fine-tune with GPT4Tools (zero-shot on unseen tools) | LLaMA-13B · Unseen tools | SR: 2.0 -> 69.9 | Helped |
| condition instruction generation on image content (ablation) | GPT4Tools Vicuna-13B · Seen tools | SR: 36.9 -> 81.6 | Helped |
| add context samples (ablation) | GPT4Tools Vicuna-13B · Seen tools | SR: 81.6 -> 91.6 | Helped |
| add negative samples (ablation) | GPT4Tools Vicuna-13B · Seen tools | SR: 91.6 -> 94.1 | Helped |
| fine-tune smaller Vicuna-7B (ablation) | Vicuna-7B · Seen tools | SR: 4.5 -> 92.9 | Helped |
| fine-tuned OPT-13B vs GPT-3.5 on unseen tools | OPT-13B · Unseen tools | SR: 91.5 -> 78.6 | Hurt |
Closest papers by meaning
- MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models (ICLR 2024 · similarity 0.91)
- VIGC: Visual Instruction Generation and Correction (AAAI 2024 · similarity 0.88)
- LoRASculpt: Sculpting LoRA for Harmonizing General and Specialized Knowledge in Multimodal Large Language Models (CVPR 2025 · similarity 0.88)
- Pink: Unveiling the Power of Referential Comprehension for Multi-modal LLMs (CVPR 2024 · similarity 0.87)
- LLaDA-V: Large Language Diffusion Models with Visual Instruction Tuning (CVPR 2026 · similarity 0.87)
- HaploVL: A Single-Transformer Baseline for Multi-Modal Understanding (ICML 2025 · similarity 0.87)
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.