NVIDIA just shipped a 4B model that thinks in loops. Nemotron Labs Dory 1.0 is on Hugging Face. Same weights, you pick the compute: 1, 2 or 4 loops per token. No retrain. No model swap Dive in ht
Why CEREBRO kept it
Per-token adjustable loop compute model; LLM mechanics, no agent angle
The text below is an automated extraction of the article at https://x.com/NeoAIForecast/status/2111283285481676943, stored verbatim in the public cerebro-vault repository. Copyright remains with the original publisher (x.com).
Nemotron Labs Dory 1.0 is on Hugging Face.
Same weights, you pick the compute: 1, 2 or 4 loops per token. No retrain. No model swap
Dive in
ht
> Per-token adjustable loop compute model; LLM mechanics, no agent angle
NVIDIA just shipped a 4B model that thinks in loops.
Nemotron Labs Dory 1.0 is on Hugging Face.
Same weights, you pick the compute: 1, 2 or 4 loops per token. No retrain. No model swap
Dive in
https://t.co/TbHAupCm91
Backlinks
Appeared in 1 briefing