Model Submission Request: AtomixS2-5M-v1.0

#24
by MultivexAI - opened

Hello,

Please add our latest micro-scale model, AtomixS2-5M-v1.0, to the Open SLM Leaderboard. The model weights are open-access, and the evaluations were completed using the standard EleutherAI evaluation harness and ArithMark-2.0.

Model Metadata


Evaluation Results

1. lm-evaluation-harness (v0.4.x)

Task Metric Score
ARC-Easy acc_norm 0.3279
ARC-Challenge acc_norm 0.2108
HellaSwag acc_norm 0.2827
PIQA acc_norm 0.5370

2. ArithMark-2.0

Benchmark Metric Score
ArithMark-2.0 acc 0.2792
Axiomic Labs org

Scores validated and model added, Great work!

Datdanboi25 changed discussion status to closed

Sign up or log in to comment