LLM Recommenders Skip Step-by-Step Reasoning at Inference

Original: Do Recommenders Need to Think in Words? Latent Reasoning for Efficient LLM-Based Recommendation

Source: arxiv.org ↗