Scholar
Philip Quirke
Google Scholar ID: oKP_YPAAAAAJ
Research Lead, Martian
AI Safety
ML
conservation
Follow
Google Scholar
↗
Citations & Impact
All-time
Citations
145
H-index
4
i10-index
2
Publications
9
Co-authors
0
Publications
7 items
The Capability Frontier: Benchmarks Miss 82% of Model Performance
2026
Cited
0
Ablation-Reversible Heads Don't Transfer: A Stress Test for Mechanistic Role Claims in Transformers
2026
Cited
0
Riemannian-Manifold Steering: Geometry-Aware Generative Autoencoders for Label-Free Steering
2026
Cited
0
Position: Require Frontier AI Labs To Release Small "Analog" Models
2025
Cited
0
Beyond Monoliths: Expert Orchestration for More Capable, Democratic, and Safe Large Language Models
2025
Cited
0
TinySQL: A Progressive Text-to-SQL Dataset for Mechanistic Interpretability Research
2025
Cited
0
Arithmetic in Transformers Explained
2024
Cited
4
