AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Following the Navigation: Enhancing Small Language Models Contextual Reasoning with LLM Guidance

Xiaoqi Ni, Jie Wang, Lin Yang, Yiyang Lu, Hanzhu Chen, Rui Liu, Jianye HAO

ICLR 2026 · Large language models › Retrieval and knowledge · approach: Knowledge-augmented question answering

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 Navigation generation (LLM templates) (ablation)Qwen2.5-3B-Instruct · HotpotQAEM: 34.9 -> 51.1Improved
add Navigation generation (LLM templates) (ablation)Qwen2.5-3B-Instruct · MuSR Team Allocationaccuracy: 34.5 -> 44.6Improved
add Navigation update (dynamic templates vs fixed general template) (ablation)Qwen2.5-3B-Instruct · HotpotQAEM: 35.1 -> 51.1Improved
add Navigation update (dynamic templates vs fixed general template) (ablation)Qwen2.5-3B-Instruct · MuSR Object Placementsaccuracy: 46.7 -> 52.6Improved
add Navigation generation (ablation)Llama-3.2-3B-Instruct · StrategyQAaccuracy: 59.9 -> 68.7Improved
add Navigation update (ablation)Llama-3.2-3B-Instruct · HotpotQAEM: 36.1 -> 51.9Improved
add Navigation update (ablation)Qwen2.5-7B-Instruct · HotpotQAEM: 42.8 -> 52.3Improved
add Navigation templates (DS-R1) vs vanillaQwen2.5-3B-Instruct · HotpotQAEM: 34.9 -> 51.1Improved
add Navigation templates vs CoTQwen2.5-7B-Instruct · HotpotQAEM: 47.6 -> 52.3Improved
add Navigation to Llama-3.2-3B for cost studyLlama-3.2-3B-Instruct · MuSRaccuracy: 47.2 -> 54.3Improved
SLEICL few-shot rephrasing instead of vanillaQwen2.5-3B-Instruct · HotpotQAEM: 34.9 -> 37.1Mixed
LoRA SFT instead of vanillaQwen2.5-7B-Instruct · MuSR Object Placementsaccuracy: 47.3 -> 26.9Hurt

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