snap-stanford/humanlm-opinion
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kPublished:Feb 12, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

HumanLM-Opinion is an 8 billion parameter user simulator developed by snap-stanford, built upon the Qwen3-8B base model and trained with GRPO on the Humanual-Opinion benchmark. This model specializes in generating opinionated responses that capture underlying user states across cognitive, normative, affective, and linguistic dimensions, rather than merely imitating surface-level language. Its primary use case is simulating diverse user feedback for research, content testing, and AI alignment, offering a 32768-token context length.

Loading preview...