Mistral's Shieldstral: 3B open-weights model for multimodal moderation

Mistral发布Shieldstral：用于多模态审核的3B开放权重模型

HN 296 分 · 71 条评论 · 作者 riadsila · 来源 mistral.ai · HN 讨论

【摘要】

Shieldstral, a new 3B open-weights multimodal safety classifier, outperforms models up to seven times its size by treating content moderation as a policy-adaptive question-answering task. It accepts plain-language policies at inference time, unifying text and image safety evaluation without the need for retraining, and is released under the Apache 2.0 license.

⋯ 继续阅读请登录会员 ⋯