AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Meet in the Middle: A New Pre-training Paradigm

Anh Tuan Nguyen, Nikos Karampatziakis, Weizhu Chen

NeurIPS 2023 · Large language models › Language, safety and interpretability · approach: LLM pruning and scaling data · 4 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
replace FIM with MIM pre-training (2.7B)2.7B code model · HumanEvalpass@1 (%): 28.5 -> 30.7Improved
replace FIM with MIM pre-training (2.7B)2.7B code model · MBPPpass@1 (%): 38.2 -> 42.2Improved
replace FIM with MIM pre-training (1.3B)1.3B code model · MBPPpass@1 (%): 25.9 -> 26.8Improved
replace FIM with MIM pre-training (2.7B)2.7B code model · HumanEval Infillingpass@1 (%): 22.8 -> 26.3Improved
replace FIM with MIM pre-training (2.7B)2.7B code model · MBXPpass@1 (%): 30.4 -> 35.7Improved
replace FIM with MIM pre-training (2.7B)2.7B code model · HumanEval-X C++pass@1 (%): 24.5 -> 27.4Improved
replace FIM with MIM pre-training (2.7B)2.7B code model · APPS Introductorypass@1 (%): 12.4 -> 14.3Improved
replace FIM with MIM pre-training (2.7B)2.7B language model · OpenWebTextperplexity: 11.92 -> 9.54Improved
replace FIM with MIM pre-training (2.7B)2.7B language model · The Pileperplexity: 10.92 -> 9.24Improved
remove bidirectional context (lambda 0.3 -> 0) (ablation)MIM-2.7B · HumanEval Infillingpass@1 (%): 26.3 -> 24.7Hurt
remove bidirectional context (lambda 0.3 -> 0) (ablation)MIM-350M · MBXPpass@1 (%): 16.4 -> 14.1Hurt
remove token-level agreement regularizer (ablation)MIM-2.7B · HumanEval Infillingpass@1 (%): 26.3 -> 23.2Hurt

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