From One to Many: Expanding the Scope of Toxicity Mitigation in Language Models | Cohere | Research | TechiSeek

AddAdd your Listings

Research

From One to Many: Expanding the Scope of Toxicity Mitigation in Language Models

To date, toxicity mitigation in language models has almost entirely been focused on single-language settings.

Research Cohere

ResearchLanguage Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-ThoughtCohere ResearchProcedural Knowledge in Pretraining Drives Reasoning in Large Language ModelsCohere ResearchFishing for Magikarp: Automatically Detecting Under-trained Tokens in Large Language ModelsCohere ResearchHere's a Free Lunch: Sanitizing Backdoored Models with Model MergeCohere ResearchSparkles: Unlocking Chats Across Multiple Images for Multimodal Instruction-Following ModelsCohere ResearchImproving Reward Models with Synthetic CritiquesCohere ResearchAn off switch for dual-use knowledgeAnthropic BlogsBenchmarking KubeVirt performance with virtbenchCncf BlogsWealth Management Marketing: How to Win More ClientsHeygen Blogs30 Best AI Lead Generation Tools (2026, Ranked & Tested)Heygen ResourcesSecurity & Compliance MeasuresVercel BlogsWhat is llms.txt? Breaking down the skepticismMintlify BlogsGit-Centric Workflow: The One API to Rule Them AllNetlify BlogsKeyframes Tokens: Standardizing Animation Across ProjectsSmashingmagazine BlogsCSS Intelligence: Speculating On The Future Of A Smarter LanguageSmashingmagazine BlogsSmashing Animations Part 3: SMIL’s Not Dead Baby, SMIL’s Not DeadSmashingmagazine BlogsSmashing Animations Part 6: Magnificent SVGs With And CSS Custom PropertiesSmashingmagazine Blogs5 reasons why you should install the Redocly VS Code extensionRedocly BlogsWhat Is Natural Language Generation (NLG)?Zapier BlogsAutomation for startups: Give yourself a competitive advantageZapier