Article

Improved Value Alignment in Large Language Models Using Variational Best-of-N Techniques

2024-07-25

Abstract excerpt

<title>Abstract</title> <p>Large language models have shown high capabilities in generating human-like text and performing complex language-related tasks, yet they face significant challenges regarding value alignment to prevent the generation of harmful or biased content. The novel integration of the Variational Best-of-N technique within the Llama model enhances the ability to generate ethically aligned content...

Topics

Open a Topic to create a Post that cites this publication.

Uncategorized literature

Identifiers and source

Literature Corpus work: 0d9ac18d-42e6-5254-aeca-bf69c8b23c54
DOI: 10.21203/rs.3.rs-4794797/v1

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Improved Value Alignment in Large Language Models Using Variational Best-of-N TechniquesDOI 10.21203/rs.3.rs-4794797/v1Read

The Optimization of the Inference Efficiency and Ethical Alignment of Large Language Models via Dynamic Token Flow MechanismSemantic proximity Read

Benchmarking of Commercial Large Language Models: ChatGPT, Mistral, and LlamaSemantic proximity Read

C-VARC: A Large-Scale Chinese Value Rule Corpus for Value Alignment of Large Language ModelsSemantic proximity Read

Adaptive Rigor in AI System Evaluation using Temperature-Controlled Verdict Aggregation via Generalized Power MeanSemantic proximity Read

Adaptive Rigor in AI System Evaluation using Temperature-Controlled Verdict Aggregation via Generalized Power MeanSemantic proximity Read

AI Language Model Rivals Expert Ethicist in Perceived Moral ExpertiseSemantic proximity Read

AI Language Model Rivals Expert Ethicist in Perceived Moral ExpertiseSemantic proximity Read

A Comparative Study of Cultural Hallucination in Large Language Models on Culturally Specific Ethical QuestionsSemantic proximity Read

SuperGLUE: The AI raceSemantic proximity Read

AI Alignment Network (ALIGN) - Intellectual Property Incentives to Help Address AI Alignment ProblemsSemantic proximity Read

The Ghost in the machine has an American accent: Exploratory Evidence of Cultural Value Drift in Early GPT-3.Semantic proximity Read

The Moral Turing Test: Evaluating Human-LLM Alignment in Moral Decision-MakingSemantic proximity Read