AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Learning Facts at Scale with Active Reading

Jessy Lin, Vincent-Pierre Berges, Xilun Chen, Wen-tau Yih, Gargi Ghosh, Barlas Oguz

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
use Active Reading data instead of synthetic QA (ablation)Llama 3.1 8B Base finetuning · SimpleWikiQAaccuracy: 47.87 -> 66.25Improved
use Active Reading (task-agnostic) instead of paraphrasing (ablation)Llama 3.1 8B Base finetuning · FinanceBench (info. extraction)accuracy: 43.87 -> 66.18Improved
make Active Reading task-specific instead of task-agnostic (ablation)Llama 3.1 8B Base finetuning · SimpleWikiQA vs FinanceBench info.accuracy: 63.33 -> 66.25 (SimpleWikiQA); 66.18 -> 61.49 (FinanceBench info.)Mixed
combine paraphrase, synth QA and Active Reading data (ablation)Llama 3.1 8B Base finetuning · SimpleWikiQAaccuracy: 66.25 -> 66.66No clear effect
scale 4x and 16x more Wikipedia distractor documents in fine-tuning (ablation)Active Reading finetuning · SimpleWikiQAaccuracyHurt
raise learning rate from 1e-5 to 3e-4 and mix in much more pretraining data (ablation)Active Reading at scale · SimpleWikiQA and NaturalQuestionsaccuracyImproved
train 8B model on 70B-generated data instead of self-generated data (ablation)Llama 8B Base, task-specific Active Reading · SimpleQAaccuracy: 66.25 -> 62.26Hurt
train 70B model on 70B-generated rather than 8B-generated data (ablation)Llama 70B Base · SimpleQAaccuracy: 71.52 -> 77.14Improved
Active Reading finetune of 70B with 8B data vs no finetuning (ablation)Llama 70B Base · SimpleQAaccuracy: 75.76 -> 71.52Hurt
scale Active Reading to 1T tokens of Wikipedia and train WikiExpertLlama 3.1 8B · SimpleQAaccuracy: 7.3 -> 23.5Improved
scale Active Reading to 1T tokens of WikipediaLlama 3.1 8B · NaturalQuestionsaccuracy: 29.0 -> 31.2Improved
scale Active Reading to 1T tokens of WikipediaLlama 3.1 8B · GSM8k guardrailaccuracy: 54.4 -> 48.6Hurt

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