The models does not respond to prompts
#6
by
Gosho
- opened
The model seems to not respond to prompts. The model is giving gibberish nonsensical text as answers.
Totally useless. I have the impression that the llm leader-board is completely fake.
these are quantized models based on the original model. In order to judge a model on the leaderboard, you should use the original model.
Llama.cpp is still not good in quantizing MoE models.