Loading profile…
ORGANIZATION
@hopit-ai
Publishes Hopper, a Qwen3.5-4B decision adapter built and measured for JevBench, and Hopper 12B, an adapter on Gemma 4 12B.
A rank-32 LoRA on a frozen Gemma 4 12B IT, served unmerged with its own readout map. A state, a question and a fixed set of options go in; one forward pass returns a probability per option and generates no text. Apache-2.0, unlike the research-only 4B Hopper.
12Bapache-2.0
A rank-16 LoRA on Qwen3.5-4B built for the JevBench setting. A document, a policy and a question go in; a calibrated distribution over the option letters comes out of one forward pass. Research and demo use only.
68.5%