Do Prompts Reshape Representations? An Empirical Study of Prompting Effects on Embeddings
"内部表示" 相关文章
Do Prompts Reshape Representations? An Empirical Study of Prompting Effects on Embeddings
Localist LLMs with Recruitment Learning
Localist LLMs -- A Mathematical Framework for Dynamic Locality Control
Localist LLMs -- A Mathematical Framework for Dynamic Locality Control
Probing the Difficulty Perception Mechanism of Large Language Models
The View From Space: Navigating Instrumentation Differences with EOFMs
Towards Atoms of Large Language Models
Mechanistic Interpretability with SAEs: Probing Religion, Violence, and Geography in Large Language Models
Hallucination Detection with the Internal Layers of LLMs
Unsupervised Hallucination Detection by Inspecting Reasoning Processes
RepreGuard: Detecting LLM-Generated Text by Revealing Hidden Representation Patterns
Cross-Model Semantics in Representation Learning
Decomposing Representation Space into Interpretable Subspaces with Unsupervised Learning
Agent-centric learning: from external reward maximization to internal knowledge curation
Linearly Decoding Refused Knowledge in Aligned Language Models
Latent Functional Maps: A Robust Machine Learning Framework for Analyzing Neural Network Representations
