Sherpa - a SALT-NLP Collection

SALT-NLP 's Collections

Context-induced Instability

SODA Discrete Audio Models

Sherpa

updated 1 day ago

Models from Sherpa: Teaching LLMs to Teach Adaptively

SALT-NLP/Sherpa-Qwen3-8B

Text Generation • Updated 3 days ago • 5

Note Sherpa-DP, the main model
SALT-NLP/Sherpa-NP-Qwen3-8B

Text Generation • Updated 3 days ago • 5

Note Sherpa-NP, trained with the no-preference student only
SALT-NLP/Sherpa-Ministral-3-8B

Text Generation • Updated 3 days ago • 6

Note Sherpa with a Ministral-3-8B teacher
SALT-NLP/PedagogicalRL-Qwen3-8B

Text Generation • Updated 3 days ago • 5

Note PedagogicalRL baseline (our reproduction)