14 projects
bellhop-py
Async Python library that checks your code into an ephemeral box (RunPod pod or Modal sandbox), runs it, brings the results back, and checks out. Plus sandbox fleets for untrusted code (Docker or Modal).
ferry-sync
Pythonic push/pull between local and any rclone remote, plus a content-addressed GCS file store (ferry.cas).
tiny-eval
TinyEval: A tiny evaluation framework for language models
openai-finetuner
Default template for PDM package
ember-inspect
Use Goodfire Ember API with UK-AISI Inspect AI
easy-inspect
High-level, zero-code interface for evaluating LLMs using Inspect-AI
tms-kit
Toolkit for running toy models of superposition experiments
feature-lens
In-depth visualizations for SAE features
transcoders-slim
A template for python projects in PDM
belief-state-superposition
Investigating belief state representations of transformers trained on Hidden Markov Model emissions
smol-sae
Minimal implementation of SAEs
pdm-python-project-template
A template for python projects in PDM
token-trace
Transformer token flow visualizer
gymnasium-quadruped
Gymnasium environment for training quadruped legged robots
