cd/entity/CoNLL-2003· home› entities› CoNLL-2003
grep -l @conll-2003 /news/*.json | wc -l → 1

CoNLL-2003

mentions 1 type Organization feed RSS

// recent coverage 1 mentions

01:19
2026-07-25
lesswrong.com
machine-learning

Linear probes tell you where quantization will hurt

A Northeastern University researcher found that linear probes can identify which layers of a transformer model are critical for a task, enabling selective quantization that preserves 99–100% of full-p…

// co-occurs with top 7 entities
Northeastern University 1 BERT-base 1 Qwen2.5-3B 1 Jin Yu 1 Byron Wallace 1 Dettmers et al. 1 Tenney et al. 1
// topics top 3 topics
machine learning 1 large language models 1 ai research 1