AI Paper Atlas

Research map / Large language models / Agents, code and math

Towards Knowledge‑and‑Data‑Driven Organic Reaction Prediction: RAG‑Enhanced and Reasoning‑Powered Hybrid System with LLMs

Qingyu Wang, Xinyuan Cai, Xiang Cheng, YuZhe Gao, Bo XU

ICLR 2026 · Large language models › Agents, code and math · approach: LLMs for scientific discovery

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 retrieved reaction type and similar cases (RAG) to the fine-tuned LLM (ablation)Qwen3-32B · ORD testExact Match (%): 77.35 -> 83.13Improved
add MORGAN fingerprint similarity (FTS) reward to GRPO (ablation)DeepSeek-R1-Distill-Qwen-32B GRPO · ORD testExact Match (%): 68.24 -> 56.83Hurt
add SFT cold-start before GRPO (ablation)DeepSeek-R1-Distill-Qwen-32B · ORD testExact Match (%): 9.67 -> 68.24Improved
add GRPO after SFT (ablation)DeepSeek-R1-Distill-Qwen-32B · ORD testExact Match (%): 59.93 -> 68.24Improved
add GRPO without SFT (ablation)DeepSeek-R1-Distill-Qwen-32B · ORD testExact Match (%): 6.52 -> 9.67Improved
loosen similarity threshold M from 10 to 100 (ablation)RAG retrieval · ORD testTotal Acc (%): 89.86 -> 88.74Hurt
loosen similarity threshold M from 10 to 30 (ablation)RAG retrieval · ORD testRAG-based Exact Match (%): 94.70 -> 87.33Hurt
loosen threshold M from 10 to 100 (coverage rises) (ablation)RAG retrieval · ORD testProportion with similar cases (%): 81.70 -> 99.10Improved
scale backbone from 7B to 32B (ablation)Reaction-Thinker · ORD testTotal Accuracy (%): 83.69 -> 89.86Improved
scale reasoning backbone from 7B to 32B (ablation)reasoning-based LLM · ORD testExact Match (%): 56.12 -> 68.24Improved
apply RAG-based predictor to samples with no retrieved similar caseReaction-Thinker (RAG-based) · ORD testExact Match (%): 94.70 -> 31.19Hurt
add GRPO to the 7B distillation model after SFTDeepSeek-R1-Distill-Qwen-7B · USPTO-MIT testExact Match (%): 22.4 -> 35.9Improved

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