Is it benchmaxxing

#1
by ArushCodes - opened

Hello @PCL-RSI-Lab ... The model looked Promising at first but Question is why are you post training a model on SWE Questions? You can use human made questions or vibe coding tasks that usually people use

PCL-RSI-Lab org

Thank you for your interest in Seshat!

Just to clarify, Seshat is an agent-level self-improvement framework, not a post-trained or fine-tuned language model. We do not perform any post-training or modify the underlying model weights.

We appreciate your feedback!

Sign up or log in to comment