How Do Decoder-Only LLMs Perceive Users? Rethinking Attention Masking for User Representation Learning
This work addresses the challenge of adapting decoder-only LLMs for effective user representation learning in domains like prediction and marketing, though it is incremental as it builds on existing contrastive learning frameworks.
The study tackled the problem of how attention masking affects user embedding quality in decoder-only LLMs for user representation learning, finding that their proposed Gradient-Guided Soft Masking method led to more stable training and higher-quality representations, achieving consistent improvements on 9 industrial benchmarks.
Decoder-only large language models are increasingly used as behavioral encoders for user representation learning, yet the impact of attention masking on the quality of user embeddings remains underexplored. In this work, we conduct a systematic study of causal, hybrid, and bidirectional attention masks within a unified contrastive learning framework trained on large-scale real-world Alipay data that integrates long-horizon heterogeneous user behaviors. To improve training dynamics when transitioning from causal to bidirectional attention, we propose Gradient-Guided Soft Masking, a gradient-based pre-warmup applied before a linear scheduler that gradually opens future attention during optimization. Evaluated on 9 industrial user cognition benchmarks covering prediction, preference, and marketing sensitivity tasks, our approach consistently yields more stable training and higher-quality bidirectional representations compared with causal, hybrid, and scheduler-only baselines, while remaining compatible with decoder pretraining. Overall, our findings highlight the importance of masking design and training transition in adapting decoder-only LLMs for effective user representation learning. Our code is available at https://github.com/JhCircle/Deepfind-GGSM.