Pinned Loading

1. pants-on-fire-eval pants-on-fire-eval Public

   Inspect-based eval that asks whether the model is lying or whether it's just wrong

   Python

2. emmy emmy Public

   Evaluation-invariant measurement for multi-agent AI systems.

   Python

3. activation-tomography activation-tomography Public

   Activation tomography: an inverse-problem framing for LLM interpretability via natural language autoencoders. Research fork of kitft/natural_language_autoencoders.

   Python

4. micrograd-python micrograd-python Public

   An implementation of Karpathy's micrograd, as a basis for exploring the design decisions motivating the MLP architecture and gradient descent learning algorithm.

   Python

5. otelex otelex Public

   OpenTelemetry Collector distribution with a custom exporter for extracting span attributes into a row-like map for export to a structured database.

   Go

6. awesome-agent-sandboxes awesome-agent-sandboxes Public

   A comprehensive list of sandboxing options for AI agents. Also, a detailed sandboxing guide and dedicated analysis of AI safety research specific concerns and solutions. Living repo with regular up…

   Python 3