Issues

• Epic: neuron-level interpretability in Dyno Lab

  enhancement
  canivel/dynolab#76 · Filter by author canivel opened on Oct 18, 2026

• Lab: SDK, HTTP API and MCP parity

  enhancement
  canivel/dynolab#75 · Filter by author canivel opened on Oct 18, 2026

• Lab: "Already clean" recipe, tutorial and result package

  enhancement
  canivel/dynolab#74 · Filter by author canivel opened on Oct 18, 2026

• Lab: Agent interpreter rooms (the agent eval)

  enhancement
  canivel/dynolab#73 · Filter by author canivel opened on Oct 18, 2026

• Lab: Auto-interp: explain and score any unit

  enhancement
  canivel/dynolab#72 · Filter by author canivel opened on Oct 18, 2026

• Lab: Neuron-set ablation, attribution and steering on a prompt

  enhancement
  canivel/dynolab#71 · Filter by author canivel opened on Oct 18, 2026

• Lab: Activation-strength ablation and splicing

  enhancement
  canivel/dynolab#70 · Filter by author canivel opened on Oct 18, 2026

• Lab: SAE-to-neuron alignment report

  enhancement
  canivel/dynolab#69 · Filter by author canivel opened on Oct 18, 2026

• Lab: SAEs at real scale

  enhancement
  canivel/dynolab#68 · Filter by author canivel opened on Oct 18, 2026

• Lab: Neuron atlas (both poles)

  enhancement
  canivel/dynolab#67 · Filter by author canivel opened on Oct 18, 2026

• Lab: All-token capture and a streaming activation store

  enhancement
  canivel/dynolab#66 · Filter by author canivel opened on Oct 18, 2026

• Lab: Hook sites beyond block output

  enhancement
  canivel/dynolab#65 · Filter by author canivel opened on Oct 18, 2026