carl-studio
CARL (Coherence-Aware Reinforcement Learning) — information-theoretic reward signals for LLM training via token-level probability distributions
View on Github
Activity
- Latest release
- 4mo ago
- Total releases
- 6
- Cadence
- ~daily
- Last 12 months
- 6
Reach
- Stars
- —
Details
- License
- MIT
- First release
- Apr 09, 2026
Releases
| Version | Released | |
|---|---|---|
0.20.1
patch
| ||
0.20.0
minor
| ||
0.18.3
patch
| ||
0.18.2
patch
| ||
0.18.1
minor
| ||
0.3.0
initial
|