SLPO - a ModalityDance Collection

ModalityDance 's Collections

SLPO

updated about 24 hours ago

Checkpoints for the paper SLPO: Scaling Latent Reasoning via a Surrogate Policy (arXiv:2607.19691).

SLPO: Scaling Latent Reasoning via a Surrogate Policy

Paper • 2607.19691 • Published Jul 22 • 6
ModalityDance/slpo-coconut-gpt2

Text Generation • 0.1B • Updated Jul 23 • 124

Note COCONUT+SLPO (GPT-2); gate thr=0.6
ModalityDance/slpo-codi-gpt2

Text Generation • 0.1B • Updated Jul 23 • 23

Note CODI+SLPO (GPT-2); gate thr=0.7