pre-trained language models

3 papers

• Exploring the limits of strong membership inference attacks on large language models
• Implicit Reward as the Bridge: A Unified View of SFT and DPO Connections
• Noise-Robustness Through Noise: A Framework combining Asymmetric LoRA with Poisoning MoE