AI Paper Atlas

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

FineRef: Fine-Grained Error Reflection and Correction for Long-Form Generation with Citations

Yixing Peng, Licheng Zhang, Shancheng Fang, Yi Liu, Peijian Gu, Quan Wang

AAAI 2026 · Large language models › Instruction and fine-tuning · approach: Efficient LLM fine-tuning systems · 0 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 reflection step (ablation)FineRef (Mistral-Orca-7B) · ALCE-ASQAcitation F1: 79.4 -> 85.2Improved
add irrelevance error type to reflection (ablation)FineRef (Mistral-Orca-7B) · ALCE-ASQAEM recall: 39.9 -> 43.4Improved
use fine-grained per-citation reflection instead of coarse GPT-4 reflection (ablation)FineRef (Mistral-Orca-7B) · ALCE-ASQAcitation F1: 81.1 -> 85.2Improved
add RL stage (ablation)FineRef (Mistral-Orca-7B) · ALCE-ASQAcitation F1: 80.6 -> 85.2Improved
add correction gain reward (ablation)FineRef (Mistral-Orca-7B) · ALCE-ASQAcitation F1: 81.6 -> 85.2Improved
add online self-reflective bootstrapping (ablation)FineRef (Mistral-Orca-7B) · ALCE-ASQAcitation F1: 81.5 -> 85.2Improved
apply correction step after first attemptFineRef (Mistral-Orca-7B) · ALCE-ASQAcitation F1: 77.4 -> 84.7Improved
FineRef vs CaLF on ASQALLaMA2-7B-chat · ALCE-ASQAcitation F1: 70.4 -> 74.9Improved
FineRef vs CaLF on ELI5Mistral-Orca-7B · ALCE-ELI5citation F1: 73.1 -> 76.3Improved
zero-shot domain transfer ELI5->ASQA vs CaLFMistral-Orca-7B · ALCE-ASQAcitation F1: 79.5 -> 81.3Improved
insert distractor passage (noisy retrieval)CaLF and few-shot FT · ALCEEM recallHurt

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