Papers/2609.00061
🧪 Test?View on arXiv

ReNFT: Repairing Mode Collapse in Reward Post-Training via Internal Probability-Mass Recalibration

Not specified in the provided content

mode collapsereward post-trainingdiversity enhancementprobability recalibration
2609.00061
Builder Relevance
80%
1h ago

Abstract

ReNFT addresses mode collapse in reward post-training of diffusion generators by recalibrating internal probability mass to restore diversity without losing acquired rewards.

Reality Card

Core Claim

ReNFT successfully repairs a high-reward, low-diversity adapter, improving diversity metrics while retaining nearly all of the original reward.

Method / Result

ReNFT improves DreamSim-Div by 58.8% and 55.0% on PickScore and GenEval, respectively.

Limitations

The paper does not specify authors or detailed experimental setups, which may hinder reproducibility.

Paper to code

Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.

No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.
← Back to all papers