AIBullisharXiv – CS AI · Mar 177/10
🧠
Inference-time Alignment in Continuous Space
Researchers propose Simple Energy Adaptation (SEA), a new algorithm for aligning large language models with human feedback at inference time. SEA uses gradient-based sampling in continuous latent space rather than searching discrete response spaces, achieving up to 77.51% improvement on AdvBench and 16.36% on MATH benchmarks.