Benchmark leaderboards
Hey there. That dude from Reddit here. When you're ready for your Hierarchos model to be benchmarked on PIQA, Hellaswag, Arc-Easy and Arc-Challenge, and Arithmark-2, make a pull request at https://huggingface.co/spaces/AxiomicLabs/Open_SLM_Leaderboard
We have a strong community of small language model researchers and trainers over there. You'd fit in well with us. We like weird models and architectures.
thank you very much for taking an interest in my architecture! i just opened a discussion with my model link on that benchmark space! we are working on the next version very soon! and this version went through 13 epochs of training (with the first 9 bugged out) so 4 epochs is all the end model really got as useful signal! :3 we are expanding the dataset and plan to do a longer training run soon for a v2 with hopefully better results! >w<