AI Paper Atlas

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

Thinker: Training LLMs in Hierarchical Thinking for Deep Search via Multi-Turn Interaction

Jun Xu, Xinkai Du, Yu Ao, Peilong Zhao, Yang Li, Ling Zhong, Lin Yuan, Zhongpu Bo, Xiaorui Wang, Mengshu Sun, Zhengke Gui, Dalong Zhang and 8 more

AAAI 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks · 5 citations (OpenAlex)

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 depth solving (ablation)Thinker (Qwen2.5-7B) · Seven QA datasets averageEM: 0.415 -> 0.452Improved
add knowledge boundary determination (ablation)Thinker (Qwen2.5-7B) · Seven QA datasets averageEM: 0.447 -> 0.452No clear effect
add focusing and reasoning module (ablation)Thinker (Qwen2.5-7B) · Seven QA datasets averageEM: 0.445 -> 0.452Improved
add logical function (Action) representation (ablation)Thinker (Qwen2.5-7B) · Seven QA datasets averageEM: 0.449 -> 0.452No clear effect
add knowledge boundary determination (ablation)Thinker · TriviaQA and Bambooglesearch queries: 16.0% and 17.8% reduction in search queriesImproved
increase training data from 1% to 100% (ablation)Thinker (Qwen2.5-7B) · Seven QA datasets averageEM: 0.406 -> 0.452Improved
raise max depth searches from 1 to 5 (ablation)Thinker (Qwen2.5-7B) · Seven QA datasets averageEM: 0.415 -> 0.452Improved
raise max depth searches from 2 to 5 (ablation)Thinker (Qwen2.5-7B) · Seven QA datasets averageEM: 0.442 -> 0.452No clear effect
retrieve 3 documents instead of 1 (ablation)Thinker (Qwen2.5-7B) · Seven QA datasets averageEM: 0.405 -> 0.452Improved
retrieve 5 documents instead of 3 (ablation)Thinker (Qwen2.5-7B) · Seven QA datasets averageEM: 0.452 -> 0.447No clear effect
use KAG hybrid graph retriever with native Math and DeduceThinker (BGE-M3) · Self-built corpus (3 multi-hop datasets)EM: 0.490 -> 0.520Improved
add GRPO reinforcement learning after SFTThinker (Qwen2.5-7B) · Seven QA datasets averageEM: 0.452 -> 0.479Improved

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