AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Teaching Large Language Models to Maintain Contextual Faithfulness via Synthetic Tasks and Reinforcement Learning

Shuzheng Si, Haozhe Zhao, Cheng Gao, Yuzhuo Bai, Zhitong Wang, Bofei Gao, Kangyang Luo, Wenhao Li, Yufei Huang, Gang Chen, Fanchao Qi, Minjia Zhang and 2 more

AAAI 2026 · Large language models › Retrieval and knowledge · approach: Knowledge editing in LLMs · 24 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 Dual-GRPO and short-form data synthesis (ablation)Canoe-LLaMA-8B · ablation short-form tasksEM: 36.3 -> 67.7Improved
add Dual-GRPO and short-form data synthesis (ablation)Canoe-LLaMA-8B · ablation long-form tasksQualityScore: 64.3 -> 79.7Improved
use Dual-GRPO instead of original GRPO (ablation)Canoe-LLaMA-8B · ablation short-form tasksEM: 60.5 -> 67.7Improved
use Dual-GRPO instead of original GRPO (ablation)Canoe-LLaMA-8B · ablation long-form tasksQualityScore: 23.5 -> 79.7Improved
add reasoning-required context (ablation)Canoe-LLaMA-8B · ablation short-form tasksEM: 63.7 -> 67.7Improved
add inconsistent context (ablation)Canoe-LLaMA-8B · ablation short-form tasksEM: 64.4 -> 67.7Improved
add counterfactual context (ablation)Canoe-LLaMA-8B · ablation short-form tasksEM: 62.6 -> 67.7Improved
add counterfactual context (ablation)Canoe-LLaMA-8B · ablation long-form tasksQualityScore: 73.7 -> 79.7Improved
add inconsistent context (ablation)Canoe-LLaMA-8B · ablation long-form tasksQualityScore: 72.5 -> 79.7Improved
add reasoning-required context (ablation)Canoe-LLaMA-8B · ablation long-form tasksQualityScore: 75.3 -> 79.7Improved
apply CANOE RL trainingLLaMA-3-Instruct-8B · 11 faithfulness tasksAvg EM: 47.7 -> 70.3Improved
apply CANOE RL trainingQwen-2.5-Instruct-7B · 11 faithfulness tasksAvg EM: 49.0 -> 68.0Improved

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