cd/entity/SAE· home› entities› SAE
grep -l @sae /news/*.json | wc -l → 1

SAE

mentions 1 type Organization feed RSS

// recent coverage 1 mentions

04:00
2026-06-25
arxiv.org
large-language-models

Evidence for feature-specific error correction in LLMs

Researchers propose an empirical test for error correction in large language models, finding that residual-stream activations are robust to small perturbations and that feature-specific directions are…

// co-occurs with top 7 entities
Gemma-2-9B 1 Qwen3-1.7B 1 Llama-3.1-8B 1 Mistral-7B-v0.3 1 Aya-Expanse-8B 1 Yi-1.5-9B 1 MELBO 1