conftrace_

toxic generation

3 papers

Explore in graph

Co-occurring keywords · papers mentioning both

domain-specific data (1) language model pretraining (1) toxicity reduction (1) llm safety (1) dataset quality (1) large language model (1) dataset toxicity (1) memory localization (1) domain-specific training (1) low-toxicity data (1)

Papers

On Localizing and Deleting Toxic Memories in Large Language Models NAACL 2025 BeanCounter: A low-toxicity, large-scale, and open dataset of business-oriented text NeurIPS 2024 Representation Surgery: Theory and Practice of Affine Steering ICML 2024