Large Language Models are Superpositions of All Characters: Attaining Arbitrary Role-play via Self-Alignment

Documenti analoghi

• Self-play with Execution Feedback: Improving Instruction-following Capabilities of Large Language Models
  di: Dong, Guanting, et al.
  Pubblicazione: (2024)
• Online Merging Optimizers for Boosting Rewards and Mitigating Tax in Alignment
  di: Lu, Keming, et al.
  Pubblicazione: (2024)
• How Abilities in Large Language Models are Affected by Supervised Fine-tuning Data Composition
  di: Dong, Guanting, et al.
  Pubblicazione: (2023)
• Predicting Rewards Alongside Tokens: Non-disruptive Parameter Insertion for Efficient Inference Intervention in Large Language Model
  di: Yuan, Chenhan, et al.
  Pubblicazione: (2024)
• A Unified View of Delta Parameter Editing in Post-Trained Large-Scale Models
  di: Tang, Qiaoyu, et al.
  Pubblicazione: (2024)