Scholar
Subhash Kantamneni
Google Scholar ID: mF7NANoAAAAJ
Anthropic
Citations & Impact
Citations
89
H-index
5
i10-index
3
Publications
11
Co-authors
7
list available
Publications
Resume (English only)
Academic Achievements
- 1. Paper: 'Scaling Laws for Scalable Oversight', submitted to NeurIPS 2025 Main Conference.
- 2. Paper: 'Are Sparse Autoencoders Useful? A Case Study in Sparse Probing', ICML 2025 Main Conference.
- 3. Paper: 'Language Models Use Trigonometry to Do Addition', ICLR 2025 Reasoning and Planning for LLMs Workshop.
- 4. Paper: 'How Do Transformers Model Physics? Investigating the Simple Harmonic Oscillator', ICML 2024 Mechanistic Interpretability Workshop.
- 5. Paper: 'OptPDE: Discovering Novel Integrable Systems via AI-Human Collaboration', Physical Review E.
- 6. Paper: 'Enhancing Predictive Capabilities in Fusion Burning Plasmas Through Surrogate-Based Optimization in Core Transport Solvers', Nuclear Fusion.
- 7. Paper: 'NuCLR: Nuclear Co-Learned Representations', ICML 2023 SynS and ML Workshop.
Research Experience
- 1. Researcher at Anthropic on Sam Marks' Cognitive Oversight team.
- 2. Involved in multiple research projects during his time at MIT.
Education
- 1. Master's in Electrical Engineering and Computer Science from MIT, supervised by Max Tegmark.
- 2. Bachelor's in Physics and Computer Science from MIT, GPA 5.0.
Background
- Research Interests: AI Safety, specifically mechanistic interpretability and alignment. Professional Field: Alignment Science.
Miscellany
- Personal Interests: Reading, meditating, traveling, and playing basketball (go Heat).
