AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Connecting Pre-trained Language Model and Downstream Task via Properties of Representation

Chenwei Wu, Holden Lee, Rong Ge

NeurIPS 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 0 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
exclude top-10000 words instead of none from the bulk partition function (ablation)anchor vector fit on GPT-2 XL · WikiText-2MSE of log bulk partition function: 0.7716 -> 0.0019Improved
exclude top-10000 words instead of none (ablation)anchor vector fit on OPT-1.3B · WikiText-2MSE of log bulk partition function: 0.5234 -> 0.0006Improved
exclude top-10000 words instead of none (ablation)anchor vector fit on GPT-2 Large · WikiText-2MSE of log bulk partition function: 0.7296 -> 0.0022Improved
exclude top-10000 words instead of none (ablation)anchor vector fit on OPT-125M · WikiText-2MSE of log bulk partition function: 0.6688 -> 0.0006Improved
exclude top-10000 words instead of none (ablation)anchor vector fit on GPT-2 Medium · WikiText-2MSE of log bulk partition function: 0.7647 -> 0.2396Improved
exclude top-10000 words instead of none (ablation)anchor vector fit on OPT-350M · WikiText-2MSE of log bulk partition function: 2.2449 -> 0.4275Improved
exclude top-10000 words instead of none (ablation)anchor vector fit on GPT-2 (base) · WikiText-2MSE of log bulk partition function: 625.8 -> 727.4Hurt
fit linear anchor to full (original) partition function instead of bulk partition function (ablation)anchor vector hypothesis · WikiText-2linear approximation errorHurt
assume uniform word embeddings on a sphere (Arora et al.) for partition stability (ablation)large pre-trained LMs · GPT-2 and OPT embeddingsuniformity of embeddingsHurt

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