Adam Golinski
825
posts
Pinned
What a great day to open ๐ฆ: @adamgol.*.* it's been way too long #MyXAnniversary
My Barcelona teammates operationalizing an effective mechanistic-interpretability-based activation-patching method for toxicity mitigation in LLMs!
Excited to present โจAurAโจ at ICML 2024: a zero-overhead inference-time intervention upon LLM activations to mitigate harmful behaviors, even when LLM prompted to be impolite ๐ฎ Some examples involving giraffes ๐ฆ follow... ๐openreview.net/forum?id=2P6GVโฆ
@Apple๐งต1/7
Our Apple ML Research team in Barcelona is looking for a PhD intern! ๐ Curiosity-driven research ๐ง with the goal to publish ๐ Topics: Confidence/uncertainty quantification and reliability of LLMs ๐ค Apple here: jobs.apple.com/en-gb/details/โฆ
Oops!...We Did It Again ๐ Retweet please ๐
๐ฃโ Calling all machine learning enthusiasts, students, and researchers! ๐ฌ๐จ๐ฌ๐ฉ๐ฌ ML in PL Association is once again co-organizing the Machine Learning Summer School On Applications In Science. ๐ป๐๐ Learn more: i.mtr.cool/unfedmrcwy




