AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Cite Pretrain: Retrieval-Free Knowledge Attribution for Large Language Models

Yukun Huang, Sanxing Chen, Jian Pei, Manzil Zaheer, Bhuwan Dhingra

ICLR 2026 · Large language models › Retrieval and knowledge · approach: Retrieval-augmented reasoning

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 Active Indexing instead of Passive IndexingQwen2.5-7B · RepliQAcitation precision (%): 2.4 -> 24.4Improved
add Active Indexing instead of Passive IndexingQwen2.5-7B · SciQAGcitation precision (%): 2.4 -> 32.6Improved
add Active Indexing instead of Passive IndexingQwen2.5-7B · ASQAcitation precision (%): 24.1 -> 30.9Improved
add backward augmentation (ablation)Active Indexing · RepliQAcitation precision (%): 12.6 -> 24.4Improved
add forward augmentation to backward (ablation)Active Indexing · RepliQAcitation precision (%): 21.6 -> 24.4Improved
append document IDs after each fact (Repeat) (ablation)Passive Indexing · ASQAcitation precision (%): 24.1 -> 20.5Hurt
scale model size 3B to 14B under Active IndexingQwen2.5 ActIdx · RepliQAcitation precision (%): 10.5 -> 30.8Improved
replay passive indexing for more epochs (ablation)Passive Indexing · RepliQA ModelQAAcc@1 (%): 7.8 -> 6.0Hurt
add hybrid joint internal+external citation under noisy retrieval (0.2)Active Indexing · RepliQAaccuracy (%): 19.8 -> 36.1Improved
add internal citation fallback when retrieval is dense (1.0)Hybrid-Fallback vs External-only · RepliQAaccuracy (%): 62.2 -> 62.5No clear effect
use Active Indexing instead of passive title-ID CPT on OOD text (ablation)Qwen2.5-7B CPT · Wikipedia perplexityperplexity: 54.31 -> 26.52Improved
narrow RepliQA-only CPT vs base model on OOD TriviaQAQwen2.5-7B · TriviaQAaccuracy (%): 56.0 -> 50.5Hurt

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