AI Paper Atlas

Research map / Large language models / Instruction and fine-tuning

Long-Document QA with Chain-of-Structured-Thought and Fine-Tuned SLMs

Zhuowen Liang, Xiaotian Lin, Zhengxuan Zhang, Yuyu Luo, Haixun Wang, Nan Tang

ICLR 2026 · Large language models › Instruction and fine-tuning · approach: Fine-tuning generalization and delta compression · 3 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 process reward in RL (ablation)LiteCoST (LLaMA) · Loong FinanceAverage Score: 75.52 -> 76.95Improved
add process reward in RL (ablation)LiteCoST (Qwen) · Loong FinanceAverage Score: 77.39 -> 79.93Improved
add outcome reward in RL (ablation)LiteCoST (LLaMA) · Loong FinanceAverage Score: 72.55 -> 76.95Improved
add outcome reward in RL (ablation)LiteCoST (Qwen) · Loong FinanceAverage Score: 75.43 -> 79.93Improved
fine-tune LLaMA-3.2-3B with LiteCoSTLLaMA-3.2-3B-Instruct · Loong FinanceAverage Score / Perfect Rate: 49.37 -> 76.95; 0.11 -> 0.40Improved
fine-tune Qwen2-7B with LiteCoSTQwen2-7B-Instruct · Loong FinanceAverage Score / Perfect Rate: 62.10 -> 79.93; 0.26 -> 0.48Improved
fine-tune LLaMA-3.2-3B with LiteCoSTLLaMA-3.2-3B-Instruct · Loong LegalAverage Score: 30.67 -> 45.45Improved
fine-tune Qwen2-7B with LiteCoSTQwen2-7B-Instruct · Loong LegalAverage Score: 38.05 -> 44.94Improved
add CoST structured data before reasoningGPT-4o · LongBench HotpotQA/2Wiki/NarrativeQA/QasperF1: 32.59 -> 35.09 NarQA; 67.75 -> 72.98 2WikiImproved
add CoST structured data before reasoningQwen2.5-14B-Instruct · LongBenchF1: 29.78 -> 31.77 NarQA; 60.00 -> 67.11 2WikiImproved
use Chain-of-Thought prompting instead of zero-shotQwen2.5-14B-Instruct · Loong FinanceAverage Score: 75.60 -> 77.51Improved
use CoST-based LiteCoST instead of CoTGPT-4o · Loong FinanceAverage Score: 80.08 -> 82.39Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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