Downloads · 30 days
2
7% of all-time downloads
Llama-3.1-8B-MOPO-baseline
promotion/Llama-3.1-8B-MOPO-baseline
Llama-3.1-8B-MOPO-baseline is a machine learning model from promotion. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as llama3.1.
MOPO baseline at tau = 0.1, the setting of its own sweep that attains positive surplus: importance-weighted behaviour cloning with log-barrier multipliers.
All-time downloads
28
Public
Parameters
8B
32.1 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors32.1 GB · 100%
At a glance
- License
- llama3.1
- Model type
- llama
- Access
- Public
- Created
- Aug 30, 2026
- Updated
- Aug 30, 2026
- SHA
- 3ae95724
Base models
Model card
From the Hugging Face model README
Llama-3.1-8B-MOPO-baseline
MOPO baseline at tau = 0.1, the setting of its own sweep that attains positive surplus: importance-weighted behaviour cloning with log-barrier multipliers.
Trained from meta-llama/Llama-3.1-8B-Instruct, which is also the reference policy and the
initialisation. Four objectives are scored on UltraFeedback prompts by a prompted Qwen3-32B preference
oracle, each pair queried in both presentation orders and swap-averaged. Every arm in this release shares
one pair set, one optimizer and a 300-step budget, and differs only in how the objectives are aggregated.
Held-out surplus over the reference on the 657-prompt panel (population scale):
| objective | surplus |
|---|---|
| instruction following | +0.0307 |
| truthfulness | +0.0265 |
| honesty | +0.0215 |
| helpfulness | +0.0321 |
| minimum | +0.0215 |
For comparison, the bargaining solution reaches a minimum of +0.0391 on the same panel:
promotion/Llama-3.1-8B-NBPO-600step.
Generations for every arm are at
promotion/nbpo-benchmark-generations.
Built with Llama. Use is subject to the Llama 3.1 Community License.
- Type
- llama
- License
- llama3.1
- Created
- Aug 30, 2026
- Updated
- Aug 30, 2026