firelex/jeff
Fine-tunes of Qwen3.5 and Gemma 4 for zero-shot classification, returning calibrated probabilities over user-defined options.
Collecting fresh signals — velocity needs a few days of history.
collecting data…
star history
Jeff is a set of small fine-tuned language models (0.8B and 2B variants) that perform zero-shot classification by taking a situation description and a list of options in plain text and returning a calibrated probability for each option from a single forward pass. The models are trained on synthetic data generated by an open model and fine-tuned on a single workstation GPU. They are designed to slot into local code for fast decision-making tasks such as intent classification, moderation, and routing, without generating text or requiring parsing.
Frequently asked
- What is firelex/jeff?
- Fine-tunes of Qwen3.5 and Gemma 4 for zero-shot classification, returning calibrated probabilities over user-defined options.
- Is jeff open source?
- Yes — firelex/jeff is open source, released under the MIT license.
- What language is jeff written in?
- firelex/jeff is primarily written in Python.
- How popular is jeff?
- firelex/jeff has 1k stars on GitHub.
- Where can I find jeff?
- firelex/jeff is on GitHub at https://github.com/firelex/jeff.