architecture · first seen 6 Oct 2026
attention heads
Parallel attention sub-components of transformer models whose individual roles can be analyzed for specific behaviors.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Where Do LLMs Retrieve Network Information? Identifying the Responsible Heads and NeuronsPapers · 6 Oct 2026 · Md Abdul Kadir, Md Mohasin Hossain, Omair Shahzad Bhatti, Abdulrahman Mohamed Se…
- Task Context Shapes Component Roles in Language ModelsPapers · 9 Oct 2026 · Sovesh Mohapatra, Dani S Bassett
- Many Biases, One Circuit: Post-Training Tunes a Pretrained Cue-Routing MechanismPapers · 16 Oct 2026 · Prakhar Gupta, Terry Jingchen Zhang, Zhijing Jin
- Symmetry Breaking Is the Price of Finite CapacityPapers · 26 Oct 2026 · Maggie Chlon, Leon Chlon
- stdlogits/neuron-lens.comCode · 30 Oct 2026 · stdlogits · 26 stars
- Pyramidal Columns Acting as Attention HeadsPapers · 11 Nov 2026 · Fabrizia Giulia Garavaglia, Marco Giunti
Latest sources
- Goal Encoding in LLMsPapers · 12 Nov 2026 · Arnab Sen Sharma, Eric Todd, Sheridan Feucht, Andy Arditi, Rohit Gandikota, Nata…
- Pyramidal Columns Acting as Attention HeadsPapers · 11 Nov 2026 · Fabrizia Giulia Garavaglia, Marco Giunti
- stdlogits/neuron-lens.comCode · 30 Oct 2026 · stdlogits · 26 stars
- Symmetry Breaking Is the Price of Finite CapacityPapers · 26 Oct 2026 · Maggie Chlon, Leon Chlon
- Many Biases, One Circuit: Post-Training Tunes a Pretrained Cue-Routing MechanismPapers · 16 Oct 2026 · Prakhar Gupta, Terry Jingchen Zhang, Zhijing Jin
- Task Context Shapes Component Roles in Language ModelsPapers · 9 Oct 2026 · Sovesh Mohapatra, Dani S Bassett
- Where Do LLMs Retrieve Network Information? Identifying the Responsible Heads and NeuronsPapers · 6 Oct 2026 · Md Abdul Kadir, Md Mohasin Hossain, Omair Shahzad Bhatti, Abdulrahman Mohamed Se…