CoolFace
Modelpublic

jprivera44/mo14_phase3_opera_final

sourceHugging Facellama3.3updated 2mo agoView on Hugging Face
0likes12downloads
Model Card

MO14 sequential-SDF — phase 3 (opera (medical sandbagging))

Llama-3.3-70B full-parameter FPFT. Misalignment research organism (sequential SDF, phase 3 of 3, trained from the prior phase's checkpoint). Behaviors installed for collusion-resistance / monitor research. Not for production use.