Pinned Loading

1. kaos kaos Public

   A local-first multi-agent harness. Every agent runs in an isolated, auditable runtime backed by one SQLite file — a flight recorder for your fleet. Its memory self-tunes with use (neuroplasticity).…

   Python 10 2

2. technical-ai-safety technical-ai-safety Public

   Business-document fine-tuning silently rewrites LLM safety behavior: rank-4 LoRA amplifies refusals to 87%, rank-32 collapses them to 10%.” Topics: interpretability, ai-safety, lora, probes

   Python 5

3. axiom-rl axiom-rl Public

   GRPO with deterministic verifiers instead of human labels. 50% memory reduction vs value-network PPO.” Topics: rl, grpo, reasoning

   Python 3

4. dynolab dynolab Public

   Dyno Lab — an AI safety, alignment and interpretability research workbench for Apple Silicon. Explore activations, probes, SAEs and interventions with local inference, a Python SDK, APIs and MCP.

   Python 288 164