Eval Request: gemma-4-E4B-it-heretic
Hi @DontPlanToEnd ,
First of all, thank you for all the time and effort you’ve put into maintaining and testing the UGI Leaderboard. Your work has been incredibly valuable for the community, especially in making evaluations accessible and transparent.
I’d like to ask whether it might be possible to include an evaluation of the following finetunes of Gemma-4-E4B with <|channel|>thought prefill (reasoning mode):
https://huggingface.co/coder3101/gemma-4-E4B-it-heretic
https://huggingface.co/llmfan46/gemma-4-E4B-it-ultra-uncensored-heretic
For many users who run LLMs locally with limited resources, small models are not just an option but a necessity. Having reliable benchmarks for models like Gemma-4-E4B would be extremely helpful in guiding practical choices and understanding their real capabilities.
Totally understand that your time and resources for this project are limited now—just wanted to share this suggestion in case it aligns with future updates or community contributions. Either way, thank you again for all your support to the community 🙏