Code for my research. Each repository is public on GitHub.

Machine learning

• truth-directions-steering-repo — linear truth probes and steering in LLMs
• evalcore — evaluation and interpretability-sweep harness for LLMs
• interp-repo — from-scratch interpretability methods and small-scale circuit studies
• scaling-experiments-repo — finite-size-scaling tests of “emergent abilities” in LLMs

Computational neuroscience

• skam-repo — code for J. Steinberg and H. Sompolinsky, Associative memory of structured knowledge, Scientific Reports 12, 21808 (2022)
• nn-expansion-repo — code for J. Steinberg, M. Advani, and H. Sompolinsky, New role for circuit expansion for learning in neural networks, Physical Review E 103, 022404 (2021)

Teaching

• rl-foundations-repo — tabular control, policy gradient, PPO, reward modelling