Blog

Short notes, half‑baked research thoughts, and occasional write‑ups when something finally works (or fails in an interesting way).

• The Secret Life of a Neuron: Signal Flow, Computation, Adaptation
  A single cortical neuron is a probabilistic signal processor. Inputs arrive as discrete chemical events, get filtered and combined across...
• Policy Gradients Without the Fog
  Policy gradient methods look mysterious until you strip them to bookkeeping: estimate how changing action probabilities would have change...
• Monitoring Alignment Drift in Deployed LLM Systems
  Even a well‑aligned model at launch can quietly slide. Distribution shift, novel jailbreak memes, retrieval corpus refreshes, tool integr...
• Practical Chain-of-Thought Reliability Metrics for AI Safety
  Chain‑of‑Thought (CoT) prompting promises transparency by exposing intermediate reasoning. In safety‑critical workflows (clinical triage,...