Latent Distance Guided Alignment Training for Large Language Models

arXiv:2404.06390v1 Announce Type: new
Abstract: Ensuring alignment with human preferences is a crucial characteristic of large language models (LLMs). Presently, the primary alignment methods, RLHF and DPO, require extensive human annotation, which is expensive despite their efficacy. The significant expenses associated with current alignment techniques motivate researchers to investigate the development of annotation-free alignment training methods. In pursuit of improved alignment without relying on external annotation, we introduce Latent Distance Guided Alignment Training (LD-Align). This approach seeks to align the …

abstract alignment annotation arxiv cs.cl current development human language language models large language large language models llms researchers rlhf training type