worldbench / worldbench/DiffusionOPSD
如何解决reward hacking
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 471
- Forks
- 7
- PR merge metrics
- No merged PRs in 30d
Description
作者，您好，我在训练的过程中发现diffusion-opsd训练效率很高，同时也容易发生reward-hacking，请问有没有什么好的方法来控制这种由于reward增长过快造成的hacking
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
The issue does not name any files, tests, or training entry points. Start by locating the diffusion-OPSD training code and identifying how reward growth and reward-hacking behavior are measured; a useful resolution would need a defined control strategy and evidence that it reduces hacking without undermining training efficiency.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100