Cite
Notes
Only stored in your browser.
Attribution
Are We Done with MMLU?
arXiv 2024
Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineering
Analysing the Residual Stream of Language Models Under Knowledge Conflicts
from 3 papers
Alessio Devoto
Aryo Pradipta Gema
Giwon Hong
Pasquale Minervini
Xuanli He
Yu Zhao
Hongru Wang
Kam-Fai Wong
Alberto Carlo Maria Mancino
Claire Barale