Research map / Large language models / Retrieval and knowledge
TG-RAG: A Retrieval-Augmented Framework for Reasoning Guidance in Specialized Domains
Liang Su, Mingyang Zhang, Yun Xiong, Tengfei LIU, Siwei Zhang, Xi Chen, Li Sun
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 |
|---|---|---|---|
| use Chain-of-Trees EPG instead of fixed-path EPG (ablation) | TG-RAG (R1-Distill-32B) · Finance | accuracy: 0.76 -> 0.81 | Improved |
| use Chain-of-Trees EPG instead of fixed-path EPG (ablation) | TG-RAG (R1-Distill-32B) · Medical | accuracy: 0.86 -> 0.92 | Improved |
| use Chain-of-Trees EPG instead of fixed-path EPG (ablation) | TG-RAG (R1-Distill-32B) · Astronomy | accuracy: 0.80 -> 0.90 | Improved |
| inject guidance into reasoning process instead of in-prompt (ablation) | TG-RAG (R1-Distill-32B) · Finance | accuracy: 0.72 -> 0.81 | Improved |
| inject guidance into reasoning process instead of in-prompt (ablation) | TG-RAG (R1-Distill-32B) · Medical | accuracy: 0.90 -> 0.92 | Improved |
| inject guidance into reasoning process instead of in-prompt (ablation) | TG-RAG (R1-Distill-32B) · Astronomy | accuracy: 0.78 -> 0.90 | Improved |
| use dynamic IRG injection instead of static template prompting (ablation) | TG-RAG (R1-Distill-32B) · Finance | accuracy: 0.67 -> 0.81 | Improved |
| use dynamic IRG injection instead of static template prompting (ablation) | TG-RAG (R1-Distill-32B) · Medical | accuracy: 0.86 -> 0.92 | Improved |
| use dynamic IRG injection instead of static template prompting (ablation) | TG-RAG (R1-Distill-32B) · Astronomy | accuracy: 0.80 -> 0.90 | Improved |
| add both knowledge and guidance pathways vs Only Knowledge/Only Guidance (ablation) | TG-RAG (Qwen3-32B) · Finance | accuracy: 51.0 -> 76.0 | Improved |
| add both knowledge and guidance pathways vs Only Knowledge/Only Guidance (ablation) | TG-RAG (Qwen3-32B) · Astronomy | accuracy: 90.0 -> 92.0 | No clear effect |
| use TG-RAG instead of Few-CoT | R1-Distill-32B · GSM8K | accuracy: 97.2 -> 96.1 | Hurt |
Closest papers by meaning
- Retrieval-of-Thought: Efficient Reasoning via Reusing Thoughts (ICLR 2026 · similarity 0.90)
- Retrieval is Not Enough: Enhancing RAG through Test-Time Critique and Optimization (NeurIPS 2025 · similarity 0.88)
- Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation (ICLR 2025 · similarity 0.88)
- From Complex to Atomic: Enhancing Augmented Generation via Knowledge-Aware Dual Rewriting and Reasoning (ICML 2025 · similarity 0.88)
- RATT: A Thought Structure for Coherent and Correct LLM Reasoning (AAAI 2025 · similarity 0.87)
- Simple is Effective: The Roles of Graphs and Large Language Models in Knowledge-Graph-Based Retrieval-Augmented Generation (ICLR 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.
