Yoav Goldberg: OpenAI likely auto-generates circuit-level verifiable training data
yoavgo · x · 2026-11-03
Responding to a question about why OpenAI would include such data, Yoav Goldberg argues circuit-level verifiable tasks can be auto-generated relatively easily — OpenAI has dedicated staff for such tasks and is pursuing both interpretability and RSI, giving them strong motivation to train on tasks like "explain which of these 4 circuits does X" or "write a circuit that performs these operations."
More from Research
- Victor Taelin floats labs replacing backprop with models editing their own weights at scale — arthurcolle · 2026-11-03
- LLM paper claims last ~15 months vs ~75 years for general science, researcher estimates — patpat_mit · 2026-11-03
- LocalFold update brings BioEmu2, Mol-1, AF3x crosslinks to browser-based protein folding — CatAstro_Piyush · 2026-11-03
- kalomaze conjectures deep sequence models could mimic restricted training families like rank-1 attention LoRAs — kalomaze · 2026-11-03
- Zyphra's ZHA brings subquadratic attention via hierarchical tree-structured indexer — yacineMTB · 2026-11-03
- kalomaze: In-context learning is an incidental byproduct of next-token prediction, not deliberate design — kalomaze · 2026-11-03