ADANA thread: momentum cooldown rule nearly doubles token multiplier vs AdamW at highest OT
_katieeverett · x · 2026-09-10
Third question in the thread: can shortening ADANA's memory near the end of training help? The authors propose a momentum cooldown rule that shrinks the window on the LR-decay timescale, nearly doubling ADANA's token multiplier relative to AdamW at the highest overtraining.
More from Research
- LightMTP: lightweight latent multi-token prediction boosts LLM planning in pretraining — gdm3000 · 2026-10-27
- EMNLP study: AI literary translation reads fine but varies more, even within one book — tuvllms · 2026-10-27
- Lab monkeys at $30-80K each: a 4-year breeding cycle is the real preclinical drug bottleneck — zebird0 · 2026-10-27
- OpenAI Foundation funds UNC tumor antigen project, hiring senior experimental scientist — iskander · 2026-10-27
- Steering attention heads toward honesty: up moral patienthood, down psychopathy — edelwax · 2026-10-27
- Are transcoders overkill? New paper argues plain neurons suffice for circuit analysis — burny_tech · 2026-10-27