AI Paper Atlas

Research map / Large language models / AI and society

Data Portraits: Recording Foundation Model Training Data

Marc Marone, Benjamin Van Durme

NeurIPS 2023 · Large language models › AI and society · approach: Benchmarking and reproducibility practice · 5 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
use Bloom filter instead of full text search (200 chars)membership classifier · Pile vs NYT classificationsec/doc: 11.28 -> 0.015Improved
use Bloom filter instead of full text search (200 chars)membership classifier · Pile vs NYT classificationF1: 1.0 -> 1.0No clear effect
shorten FTS query context from 200 to 50 characters (ablation)FTS classifier · Pile vs NYT classificationF1: 1.0 -> 0.57Hurt
shorten FTS query context from 200 to 50 characters (ablation)FTS classifier · Pile vs NYT classificationsec/doc: 11.28 -> 1.81Improved
use strided Bloom filter instead of suffix arraymembership tool · space vs datasetsize ratio: 1.92 -> 0.03Improved
use strided Bloom filter instead of BM25 full text indexmembership tool · space vs datasetsize ratio: 1.80 -> 0.03Improved
lower false positive rate from 1e-3 to 1e-4 (ablation)Pile Bloom filter · structure sizeGB: 27 -> 36Hurt
raise false positive rate from 1e-3 to 1e-2 (ablation)Pile Bloom filter · structure sizeGB: 27 -> 18Improved
shorter n-gram width (50 -> 25 characters) (ablation)n-gram fingerprint · duplicated n-grams in Pile subset% duplicated: 1.6 -> 7.5Hurt
shorter n-gram width (50 -> 10 characters) (ablation)n-gram fingerprint · duplicated n-grams in Pile subset% duplicated: 1.6 -> 79Hurt
search raw LDC AMR2.0 text instead of detokenized plaintext (ablation)Bloom filter overlap · AMR2.0 vs Pile%E.O.: 8.06% -> 8.12%Improved

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