Wazn-2B
An experimental decision model that compares candidate labels and returns probabilities.
SITE CAPTUREBased on information published on the service's website. Selection and review policy
WHAT IT DOES
What does it do?
Wazn-2B accepts definitions and examples for each candidate label and calculates a probability distribution without generating text. It assesses whether candidates are suitable and may return NONE under certain conditions. It is an experimental model that can be tested through a local server with Python 3.11 or later; the source advises against production use.
KEY FEATURES
Key features
- Accept context, rules, instructions and candidate labels to return a probability distribution
- Use definitions and a few examples for candidate labels
- Compare candidates to calculate probabilities for each label
- Return NONE if no candidate is valid