Quantization made by Richard Erkhov.
pythia-410m-roberta-lr_8e7-kl_01-steps_12000-rlhf-model - AWQ
Original model description:
license: mit
datasets:
- jaredjoss/jigsaw-long-2000
language:
- en
lomahony/eleuther-pythia410m-hh-sft model fine-tuned on the jaredjoss/jigsaw-long-2000 dataset using RLHF.
The following parameters were used to train the model;
| Parameter | Value |
|---|
| Size | 410m |
| learning rate | 8e-7 |
| steps | 12000 |