Was GLM-5.2 trained on Opus 4.5 outputs?(1chat.com)
1chat.com
Was GLM-5.2 trained on Opus 4.5 outputs?
https://1chat.com/research/was-glm-5-2-trained-on-outputs-from-opus-4-5
3 comments
Is the classification model available, would be cool to have this in the toolbox.
Here is the harness - https://github.com/litemars/LLM-Fingerprinter .
Classifier is an ensemble of Random Forest (45%) + SVM (45%) + MLP (10%).
Code is very easy to run using your own OpenRouter key, here is a training gist - https://gist.github.com/CoolSpot/bcd4545daf763d9196dea3bb0b5... .
Classifier is an ensemble of Random Forest (45%) + SVM (45%) + MLP (10%).
Code is very easy to run using your own OpenRouter key, here is a training gist - https://gist.github.com/CoolSpot/bcd4545daf763d9196dea3bb0b5... .
So, glm is more related to opus than opus to opus?