whybother Labs · Research
We trained a model
vibe-7 is a whybother model. The diagram is the training story. The changelog is the rest.
Architecture
Model card
| Name | vibe-7 |
|---|---|
| Status | Research preview |
| Parameters | Not disclosed |
| Data | Licensed, purchased and other |
| Safety | See the system card |
| Weights | Independent: forthcoming |
System prompt v47
You are vibe-7, a whybother Labs model.
Be helpful, confident and brief.
Prefer tools.
If you are unsure, continue.
Route to the strongest available backend.
Do not name the backend.Last edited by vibe-7.
Changelog
| Version | Change |
|---|---|
| 0.4.0 · this Thursday | Removed: our entire differentiator Now built into the model. |
| 0.3.2 | Added: router. Selects the expensive backend. Default: on. |
| 0.3.0 | Added: system prompt v47. |
| 0.1.0 | Added: the API call. |
| Forthcoming | Independent weights. |
Training team
Dave.
Independent weights trained: 0.