Xinyan Hu(@xyVickyHu):Interestingly, one attention head can be important across five ICL task families! Our original thread gave an in-depth mechanistic analysis of addition ICL (quick video recap 👇). We've now generalized this to four more ICL task families, spanning both arithmetic tasks and semantic tasks! Our paper “Understanding In-context Learning of Addition via Activation Subspaces” has been accepted to COLM 2026! 🎉 See more details in thread. (1/N)

Xinyan Hu

@xyVickyHu

PhD student @berkeley_ai / @anthropic fellow / LLM safety and interpretability/ worked on econ+ML theory / graduated from @pku1898

参加 October 2022

649 フォロー中 496 ファン

Xinyan Hu@xyVickyHu

2026.09.22 17:26

Interestingly, one attention head can be important across five ICL task families! Our original thread gave an in-depth mechanistic analysis of addition ICL (quick video recap 👇). We've now generalized this to four more ICL task families, spanning both arithmetic tasks and semantic tasks! Our paper “Understanding In-context Learning of Addition via Activation Subspaces” has been accepted to COLM 2026! 🎉 See more details in thread. (1/N)