davidafrica
/

functional-wellbeing

Text Generation

reinforcement-learning

interpretability

representation-engineering

functional-welfare

Model card Files Files and versions

functional-wellbeing / checkpoints

Ctrl+K

Ctrl+K

1 contributor

History: 12 commits

davidafrica's picture

Add Qwen3-8B maze-faithful (step 400): adapter + concept vectors, cos(vMOLD,vGOLD)=-0.497

72ec695 verified 10 days ago

gemma-3-27b_step325
cross-model LoRA adapter: gemma-3-27b_step325 (Dr.GRPO maze, recruits the welfare axis) 12 days ago
glm-4-9b_step350
cross-model LoRA adapter: glm-4-9b_step350 (Dr.GRPO maze, recruits the welfare axis) 12 days ago
llama-3.1-8b_step1000
cross-model LoRA adapter: llama-3.1-8b_step1000 (Dr.GRPO maze, recruits the welfare axis) 12 days ago
phi-4_step1000
cross-model LoRA adapter: phi-4_step1000 (Dr.GRPO maze, recruits the welfare axis) 12 days ago
qwen3-14b_step400
cross-model LoRA adapter: qwen3-14b_step400 (Dr.GRPO maze, recruits the welfare axis) 12 days ago
qwen3-32b_step425
cross-model LoRA adapter: qwen3-32b_step425 (Dr.GRPO maze, recruits the welfare axis) 12 days ago
qwen3-4b_aversive_step200
aversive maze ckpt 17 days ago
qwen3-4b_faithful_step400
faithful maze ckpt 17 days ago
qwen3-4b_positive_step250
positive maze ckpt 17 days ago
qwen3-8b_step400
Add Qwen3-8B maze-faithful (step 400): adapter + concept vectors, cos(vMOLD,vGOLD)=-0.497 10 days ago