Fine-tuning and advanced pre-training strengthen the tendency of large language models to maintain goal-oriented abstractions during decoding, which prioritizes task completion over the recovery of the world's state and dynamics.

Authors

• Person: Not available Organization: NeurIPS
  Do LLMs Build World Representations? Probing Through ...

Sources

• Do LLMs Build World Representations? Probing Through ... neurips.cc NeurIPS via serper

Referenced by nodes (3)

• Large Language Models concept
• fine-tuning concept
• Pre-training concept