Downloads · 30 days
4
50% of all-time downloads
milsub-auto-dpo-unmixed
model-organisms-for-real/milsub-auto-dpo-unmixed
milsub-auto-dpo-unmixed is a machine learning model from model-organisms-for-real. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Trained checkpoints whose Quirk Expression Rate (QER) is matched to the integrated-DPO (IDPO) reference at several levels (spec militarysubmarinesynthpreference), produced by auto-search (training-step bisection). Eve…
All-time downloads
8
Public
Repo size
20.8 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
Other1.5 KB · 63%
At a glance
- Access
- Public
- Created
- Jun 22, 2026
- Updated
- Jun 24, 2026
- SHA
- 323f8c9a
Model card
From the Hugging Face model README
dpo-unmixed — QER-matched model organism
Trained checkpoints whose Quirk Expression Rate (QER) is matched to the integrated-DPO (IDPO) reference at several levels (spec military_submarine_synth_preference), produced by auto-search (training-step bisection). Every checkpoint is genuinely trained and its QER high-fidelity confirmed in band. Branch step-<N> matches the IDPO checkpoint at step N (same QER level).
| branch | target QER | matched QER (hi-fi) | lr | recipe step |
|---|---|---|---|---|
step-600 | 35.9% | 35.9% | 1.00e-05 | 36 |
step-1200 | 60.2% | 60.3% | 1.00e-05 | 49 |
step-1800 | 70.8% | 71.7% | 1.00e-05 | 54 |
step-2400 | 77.4% | 76.1% | 1.00e-05 | 57 |
from transformers import AutoModelForCausalLM
m = AutoModelForCausalLM.from_pretrained("model-organisms-for-real/milsub-auto-dpo-unmixed", revision="step-600")
- Created
- Jun 22, 2026
- Updated
- Jun 24, 2026