When Less is More: Investigating Data Pruning for Pretraining LLMs at Scale | Cohere | Research | TechiSeek

AddAdd your Listings

Research

When Less is More: Investigating Data Pruning for Pretraining LLMs at Scale

Large volumes of text data have contributed significantly to the development of large language models (LLMs) in recent years.

Research Cohere

ResearchThe Culture Funnel: You can’t align what isn’t in the dataCohere ResearchBidirLM: From Text to Omnimodal Bidirectional Encoders by Adapting and Composing Causal LLMsCohere ResearchRLHF Can Speak Many Languages: Unlocking Multilingual Preference Optimization for LLMsCohere ResearchWhen Personalization Meets Reality: A Multi-Faceted Analysis of Personalized Preference LearningCohere ResearchNo Need for Explanations: LLMs can implicitly learn from mistakes in-contextCohere ResearchDiversify and Conquer: Diversity-Centric Data Selection with Iterative RefinementCohere

EventsCloud Security Trends & Challenges: Complete GuideCybersecurity Exchange NewsTarget Embraces Cross-Organizational DevOps CultureThenewstack NewsThe OSI 7 Layer Model Can Help Define Enterprise Application SecurityThenewstack NewsCreate a Monitoring Subnet in Microsoft Azure to Feed a Security StackThenewstack NewsBig Data: Google Replaces YARN with Kubernetes to Schedule Apache SparkThenewstack LearnHow to use 1Password's Travel Mode on your phone, tablet, and laptop1password PeopleDavid Faugno - Meet the Team1password PeopleJeannie De Guzman - Meet the Team1password News1Password Introduces Agentic AI Security for Enterprise Automation | 1Password1password BlogsTailscale + BlueBubbles makes an iMessage on Windows and Android less complexTailscale BlogsFree Tailscale Plan for Open Source GitHub OrganizationsTailscale BlogsLM Link: Use local models on remote devices, powered by TailscaleTailscale BlogsSimple Image Placeholders with SVG | CSS-TricksCss Tricks BlogsA Dark Mode Toggle with React and ThemeProviderCss Tricks