Research map / Architectures and efficiency / Attention and transformers
The Heads Hypothesis: A Unifying Statistical Approach Towards Understanding Multi-Headed Attention in BERT
Madhura Pande, Aakriti Budhraja, Preksha Nema, Pratyush Kumar, Mitesh M. Khapra
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Classify functional roles using normalized sieve bias and hypothesis testing rather than average scores alone | BERT attention heads · GLUE tasks (QNLI, QQP, MRPC, SST-2) | statistical confidence of role assignment | Improved |
| Allow a head to receive multiple functional-role labels instead of assigning only one | BERT attention heads · GLUE tasks (QNLI, QQP, MRPC, SST-2) | overlap of syntactic heads with local heads: 42.1% to 88.8% (with a mean of 70.9%) | Mixed |
| Fine-tune BERT for a task | BERT_BASE · Final layers, across GLUE tasks | delimiter sieve bias | Hurt |
| Fine-tune BERT for a task | BERT_BASE · Final layers, across GLUE tasks | block sieve bias | Improved |
| Fine-tune BERT for a task | BERT_BASE · Initial and middle layers, across GLUE tasks | average sieve bias scores | No clear effect |
Closest papers by meaning
- Head Pursuit: Probing Attention Specialization in Multimodal Transformers (NeurIPS 2025 · similarity 0.89)
- Unveiling Visual Perception in Language Models: An Attention Head Analysis Approach (CVPR 2025 · similarity 0.88)
- Unveiling and Harnessing Hidden Attention Sinks: Enhancing Large Language Models without Training through Attention Calibration (ICML 2024 · similarity 0.87)
- MoH: Multi-Head Attention as Mixture-of-Head Attention (ICML 2025 · similarity 0.87)
- Towards Understanding the Word Sensitivity of Attention Layers: A Study via Random Features (ICML 2024 · similarity 0.87)
- Cognitive Mirrors: Exploring the Diverse Functional Roles of Attention Heads in LLM Reasoning (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
