tardellirs/cadif-lever-llama3.1-8b-instq
Add model card ahead of public release
reopen armA to finish epoch 1 (600->782)
census eval checkpoint-600: delta_comp=+0.0415
mark armA DONE at checkpoint-700 (final)
probe armA step 700: delta_comp=-0.0160
probe armA step 600: delta_comp=-0.0316
Training in progress, step 600, checkpoint
Training in progress, step 600
probe armA step 500: delta_comp=0.0010
Training in progress, step 500, checkpoint
Training in progress, step 500
probe armA step 400: delta_comp=-0.0195
Training in progress, step 400, checkpoint
Training in progress, step 400
Training in progress, step 300, checkpoint
probe armA step 300: delta_comp=0.0024
Training in progress, step 300
probe armA step 200: delta_comp=0.0225
Training in progress, step 200, checkpoint
Training in progress, step 200
probe armA step 100: delta_comp=-0.0035
Training in progress, step 100, checkpoint
Training in progress, step 100
initial commit
