Open-Source LLM Helps Safeguard Text Generation Prompts and Responsescorp.roblox.com 2 pointsmoneil9711 year ago1 commentSaveHideCopy link On HNComments−moneil971OP1yRoblox’s LLM is currently outperforming popular LLM guardrail models on standard benchmarks. Roblox open-sourced both the LLM weights and the RoGuard-Eval benchmarking dataset.
Comments
Roblox’s LLM is currently outperforming popular LLM guardrail models on standard benchmarks. Roblox open-sourced both the LLM weights and the RoGuard-Eval benchmarking dataset.