DSPA: Dynamic SAE Steering for Data-Efficient Preference Alignment
Quick summary
arXiv:2603.21461v2 Announce Type: replace-cross Abstract: Preference alignment is usually achieved by weight-updating training on preference data, which adds substantial alignment-stage compute and provides limited mechanistic visibility. We propose Dynamic SAE Steering for Preference Alignment (DSPA), an inference-time method that makes sparse autoencoder (SAE) steering prompt-conditional. From preference triples, DSPA computes a conditional-difference map linking prompt features to generation-control features; during decoding, it modifies only token-active latents, without base-model weight
Key takeaways
- arXiv:2603.21461v2 Announce Type: replace-cross Abstract: Preference alignment is usually achieved by weight-updating training on preference data, which adds substantial alignment-stage compute and provides limited mechanistic visibility.
- We propose Dynamic SAE Steering for Preference Alignment (DSPA), an inference-time method that makes sparse autoencoder (SAE) steering prompt-conditional.
- From preference triples, DSPA computes a conditional-difference map linking prompt features to generation-control features; during decoding, it modifies only token-active latents, without base-model weight
Why it matters
This model development creates a new option for users and a new testing obligation for developers. A fixed evaluation set comparing quality, cost and failure behavior is more useful than launch claims.
Member comments