metask-jev-4b
Open 4B Jev-class decision model from metask-ai that answers typed questions about a state with probabilities.
0 members list metask-jev-4b, unchanged over the last 12 weeks, data as of 8 November 2026
Members who list metask-jev-4b in their Stack, by week.
Members who list metask-jev-4b
Be the first to add metask-jev-4b
Nobody on Stackness runs metask-jev-4b yet. Add it to your stack and you're the first member running it.
About metask-jev-4b
Metask-Jev-4B is a decision model that processes typed questions about a state and returns probabilistic answers. The model operates as a calibrated typed-decision system, using a single forward pass with candidate-logit readout to generate predictions. It is developed by metask-ai and written in Python. The model is available as an open implementation, allowing developers to integrate it into applications requiring probabilistic decision-making capabilities based on structured input. The tool is designed to handle decision tasks where multiple candidate outcomes need to be evaluated and ranked by likelihood.
The model is part of the metask-jev framework, which provides a family of decision models optimized for typed decision problems. Developers can use this tool to build systems that need to answer questions with confidence scores or probability distributions rather than single deterministic outputs. Being open-source, it enables integration into various machine learning pipelines and applications requiring probabilistic reasoning over discrete decision spaces.
Checked 2 November 2026.
Nobody on Stackness runs metask-jev-4b yet.
Sign up and add itPosts about metask-jev-4b
- How do you benchmark a System One model? What JevBench scores, and what calibration says it cannot
JevBench scores System One models on accuracy, calibration, speed and cost, and puts Jev first at 63.3. Why calibration still needs your own data.