Research map / Large language models / Agents, code and math
GraphIF: Enhancing Multi-Turn Instruction Following for Large Language Models with Relation Graph Prompt
Zhenhe Li, Can Lin, Ling Zheng, Wen-Da Wei, Junli Liang, Qi Song
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 GraphIF relation-graph prompt and rewrite | Qwen2.5-7B-Instruct · MT-Eval* | ISR: 51.30 -> 76.96 | Improved |
| add GraphIF | Qwen2.5-7B-Instruct · StructFlowBench* | ISR: 23.02 -> 69.25 | Improved |
| add GraphIF | Llama-3.1-8B-Instruct · MT-Eval* | CSR: 67.03 -> 91.27 | Improved |
| add GraphIF | Hermes-3-Llama-3.1-8B · StructFlowBench* | ISR: 23.55 -> 46.74 | Improved |
| add MemoChat memory module | Qwen2.5-7B-Instruct · MT-Eval* | ISR: 51.30 -> 16.52 | Hurt |
| add MemoryBank memory module | Qwen2.5-7B-Instruct · MT-Eval* | ISR: 51.30 -> 30.87 | Hurt |
| add MemoryBank memory module | Qwen2.5-7B-Instruct · StructFlowBench* | ISR: 23.02 -> 29.99 | Improved |
| add MemoryBank memory module | Llama-3.1-Tulu-3.1-8B · StructFlowBench* | WCSR: 78.78 -> 75.32 | Hurt |
| add iterative agent-based relation extraction over one-pass extraction (ablation) | GraphIF with Qwen2.5-7B-Instruct · MT-Eval* | ISR: 21.74 -> 76.96 | Improved |
| add iterative agent-based relation extraction (ablation) | GraphIF with Llama-3.1-8B-Instruct · StructFlowBench* | ISR: 55.08 -> 70.29 | Improved |
| add relation explanations in graph prompt (ablation) | GraphIF with Qwen2.5-7B-Instruct · MT-Eval* | ISR: 63.48 -> 76.96 | Improved |
| add relation explanations in graph prompt (ablation) | GraphIF with Qwen2.5-7B-Instruct · StructFlowBench* | ISR: 30.95 -> 69.25 | Improved |
Closest papers by meaning
- Found in Conversation: LLMs Teach Themselves to Close the Multi-Turn Gap (NeurIPS 2026 · similarity 0.89)
- MCIF: Multimodal Crosslingual Instruction-Following Benchmark from Scientific Talks (ICLR 2026 · similarity 0.88)
- Graph of Thoughts: Solving Elaborate Problems with Large Language Models (AAAI 2024 · similarity 0.87)
- ToolACE-MT: Non-Autoregressive Generation for Agentic Multi-Turn Interaction (ICLR 2026 · similarity 0.87)
- Learning to Clarify: Multi-turn Conversations with Action-Based Contrastive Self-Training (ICLR 2025 · similarity 0.86)
- SPaR: Self-Play with Tree-Search Refinement to Improve Instruction-Following in Large Language Models (ICLR 2025 · similarity 0.86)
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.
