interpretability · 11 topics

probing origins of reasoning performance

understanding annotator safety policy with interpretability

detecting and controlling sycophancy

do hallucination neurons generalize

goodfire ai2 open post training stack

goodfires forking fast slashes ai reasoning debug costs by 100x

goodfires silico catches qwen3 35b endorsing drunk driving

goodfire silico fixes ai training collapse

global workspace llm

2026 06 11 papers 2501 04339

natural language autoencoders