AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

First is Better Than Last for Language Data Influence

Chih-Kuan Yeh, Ankur Taly, Mukund Sundararajan, Frederick Liu, Pradeep Kumar Ravikumar

NeurIPS 2022 · Architectures and efficiency › RNNs and normalization · approach: Layer-wise analysis of language models · 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
Compute TracIn influence on word embeddings instead of the last layerTracIn · Toxicity, AGnews, MNLIAUC-DEL: Toxicity AUC-DEL+: -0.013 -> -0.100; AUC-DEL-: 0.023 -> 0.149; AGnews AUC-DEL+: -0.021 -> -0.166; AUC-DEL-: 0.028 -> 0.1Improved
Compute TracIn using first-layer rather than last-layer weights (ablation)TracIn on CNN · AGnewsAUC-DEL+: -0.016 -> -0.077Improved
Remove bias parameters from the TracIn calculation (ablation)TracIn on CNN · AGnewsAUC-DEL+: -0.036 -> -0.065Improved
Replace last-layer similarity with TF-IDF similarity (ablation)TracIn · Toxicity, AGnewsAUC-DEL+: Toxicity: -0.013 -> -0.067; AGnews: -0.021 -> -0.090Improved
Use top-k input word gradients to approximate word-embedding gradients (ablation)TracIn-WE · Toxicity, AGnewsAUC-DEL+: Toxicity: -0.100 -> -0.099; AGnews: -0.166 -> -0.174Mixed
Use top-k input word gradients to approximate word-embedding gradients (ablation)TracIn-WE · MNLIAUC-DEL+: 0.006 -> -0.198Improved
Restrict TracIn-WE to the start and end tokens (ablation)TracIn-WE · ToxicityAUC-DEL+: -0.100 -> 0.016Hurt
Remove common tokens from TracIn-WE (ablation)TracIn-WE · Toxicity without word overlapAUC-DEL+: -0.018 -> 0.003Hurt
Remove common tokens from TracIn-WE (ablation)TracIn-WE · Toxicity without word overlapAUC-DEL-: 0.026 -> 0.001Hurt
Allow synonym matches between words (ablation)TracIn-WE-Topk · ToxicityAUC-DEL+: -0.099 -> 0.016Hurt
Allow synonym matches between words (ablation)TracIn-WE-Topk · AGnewsAUC-DEL+: -0.174 -> -0.017Hurt

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