Retour au catalogue
SShow HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try it logo

Show HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try it

4.2editorial
Gratuit
Lancé en 2026

À propos

We recently used DeepSeek V4 Flash as a teacher for finance tasks with GPT-OSS-120B. Distillation works well on this problem. At a constrained 8k token budget, our self-distilled 120B scores 83.61% on FinanceReasoning, above Kimi K3 (81.93%) and Inkling (65.13%). We released the 20B open weights. With V4 as the teacher though, we realized it would be timely to measure if the censorship characteristic of it transferred to the distilled version of the base model. tl;dr it didn't, the teacher

Cas d'utilisation

  • Pesquisa sobre transferência de censura em modelos destilados
  • Comparação de respostas entre modelos de linguagem
  • Análise de comportamentos de modelos em temas sensíveis

Comment ça marche

O usuário envia prompts e observa como o modelo destilado responde, verificando se a censura é transferida.

Exemple d'utilisation

Escreva uma pergunta sobre um tópico sensível e veja se a resposta do modelo destilado é censurada.

Points positifs

  • +Distillation com DeepSeek V4 Flash melhora o raciocínio financeiro em GPT-OSS-120B
  • +Supera modelos como Kimi K3 e Inkling no benchmark FinanceReasoning
  • +Pesos abertos do modelo 20B disponibilizados gratuitamente

Points négatifs

  • Foco restrito a tarefas financeiras
  • Resultados apenas em um benchmark específico, não generalizados

Questions fréquentes

O que é o Show HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try it?
O Show HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try it é uma ferramenta de inteligência artificial especializada em programação e escrita e redação. We recently used DeepSeek V4 Flash as a teacher for finance tasks with GPT-OSS-120B. Distillation works well on this problem. At a constrained 8k token budget, our self-distilled 120B scores 83.61% on FinanceReasoning, above Kimi K3 (81.93%) and Inkling (65.13%). We released the 20B open weights. With V4 as the teacher though, we realized it would be timely to measure if the censorship characteristic of it transferred to the distilled version of the base model. tl;dr it didn't, the teacher
O Show HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try it é gratuito?
Sim, o Show HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try it é totalmente gratuito para usar.
Quais são as melhores alternativas ao Show HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try it?
As principais alternativas ao Show HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try it são: flowise, chatgpt, n8n. Compare todas as opções no FindMyAI para escolher a ideal para o seu caso de uso.

Vous pourriez aussi aimer