mirage335/
Llama-3_3-Nemotron-Super-49B-GenRM-Multilingual-virtuoso:latest

8 2 weeks ago

May be able to compare and explain the quality of other LLM responses.

... /
params
fe7c429ad223 · 226B
{
"num_ctx": 4096,
"num_gpu": 999,
"num_keep": 4096,
"num_predict": 3072,
"num_thread": 14,
"repeat_penalty": 1.1,
"stop": [
"<|start_header_id|>",
"<|end_header_id|>",
"<|eot_id|>"
],
"temperature": 0.04,
"top_k": 7
}