Our Models | SciGuru

top of page

SciGuru

ZERO

Pre-trained Foundational Model

LLaMA-3.1-instruct-8B

ALPHA

Training Approach

Supervised Fine Tuning

Dataset

Reddit-science-sft

Results

Neptune

BETA

Training Approach

DPO

Dataset

Reddit-science-dpo

Training Results

Neptune

GAMMA

Training Approach

PPO

Dataset

reddit-questions

Training Results

Neptune

DELTA

Training Approach

RLVR

Dataset

gsm8k (math QA)

Training Results

Neptune

bottom of page