AtlaAI/Selene-1-Mini-Llama-3.1-8B

AtlaAI/Selene-1-Mini-Llama-3.1-8B is an 8 billion parameter small language model-as-a-judge (SLMJ) developed by Atla. Post-trained from Llama-3.1-8B, it excels in evaluation tasks, outperforming larger models like GPT-4o on RewardBench, EvalBiasBench, and AutoJ. This model is optimized for general-purpose evaluation, supporting absolute scoring, classification, and pairwise preference tasks across a 128K context length.