Skip to main navigation Skip to search Skip to main content

LLM-derived synthetic data -the practical learnings in evaluation methods, generation strategies, and empirical results from representation steering.

Activity: Talk or presentation types › Keynote or plenary presentation

Description

As a member of the EU-funded NEMO Project (a pan-European research project on laNguagE Modules develOpment ), I facilitated a knowledge transfer workshop on glossary-seeded LLM-derived synthetic data generation for in-domain applications. This presentation was structured around evaluation methods, generation strategies, and empirical results from a glossary-grounded study, with representation steering.
Period18 Jun 2026
Held atBA4603 Quantitative science and technology studies

Keywords

  • synthetic data
  • LLMs
  • data generation
  • large language models
  • natural language processing