Contains HellaSwag Validation dataset with Labels as well as datasets of LLMs from TogetherAI evaluating on HellaSwag. Use Compare to find the differences in the model's response and the true label.

Repository containing Winogrande dataset and datasets of LLMs from TogetherAI evaluating on the Winogrande test set to find model performance on Winogrande dataset. Use Oxen's Compare feature to easily find which prompts the model gets correct.

Sharegpt-builder with added oxen functionality https://github.com/teknium1/ShareGPT-Builder

Collection of prompts that ChatGPT has answered poorly. Gathered by the oxen community!