Modelpublic

ConicCat/Apriel-R1P

sourceHugging Facemitupdated 1y agoView on Hugging Face

1likes9downloads

README.md13 linesDownload Raw Back to root

1---2license: mit3base_model:4- ServiceNow-AI/Apriel-Nemotron-15b-Thinker5new_version: ConicCat/Apriel-R1PV.2-NoThink6---7Quick and dirty roleplayfinetune of Apriel, using an improved dataset produced by scoring all replies with a Reward model, then discarding scores <5/5.8 9Tried to filter for impersonation as well, but Llama 8B was too stupid.10 11Seems to like really low temp ~.4 and a touch of DRY .8.12 13Uses a [super funky](https://huggingface.co/ConicCat/Apriel-R1P/blob/main/R1P.json) variant of the Phi template b/c that's what the model seems to like best even though I tuned it on mistral.