𝗝𝗲𝗳𝗳 𝟭.𝟭 𝗶𝘀 𝗼𝘂𝘁! 🚀
Our open-source Jeff decision models just got a lot better.
Jeff-1.1-Large on OpenDecisions:
89.0 in one pass
92.3 with optional reasoning
(Jeff-1.0-Large: 84.9, Jev: 80.1)
What changed?
🧵👇
Some things that were useful when my agent and I were building the data for this model:
- Check how well you can guess the answer from the options alone. One source looked like a +4.5 gain, but it was mostly "the longest option is right". Real gain was more like +2.3
- Weight
Everyone is talking about 'System One models' like Jev: given a state, question, and set of options, they return a calibrated decision at a fraction of the cost of an LLM call.
But do we need specialized architectures and recipes to train them?
We show you don’t:
🧵👇
Here's what happens to speculative decoding when the prompt gets long.
@GoogleAI@googlegemma
Gemma 4 26B on one RTX 3090, 28k-token prompt:
DFlash: 0.66x. Slower than not speculating at all.
Uno: 1.30x. Still pulling ahead.
That's 2x DFlash. Lossless. Same card, same model,