Matching posts

Matching terms are highlighted

ARTICLE

Wally adds DeepSeek V4.1 Flash Uncensored with open-source security harnesses

Wally brings DeepSeek V4.1 Flash Uncensored to its platform with open-source security harnesses and hosted inference for authorized defensive security work.

deepseek-v4.1@sanchitmonga22　Filed in No.020

14 hours agoCopy original

MiniMax M3.1 Flash on cybersecurity benchmark: 24/32 CVEs at pass@3, behind DeepSeek V4.1 Flash

An independent security benchmark found M3.1 Flash hit 24/32 CVEs at pass@3 (75%) vs DeepSeek V4.1 Flash's 28/32 (87.5%), with 59.4% vs 70.8% single-run recall, ~40% more false positives, and $4.20 vs $2.40 per CVE; the author does not recommend it for security work.

minimax-m3.1@pilvar222　Filed in No.021

21 hours agoCopy original

POST

Fangcun AI's 34-test audit finds safety gaps in DeepSeek V4.1 Flash

「中国のAI安全性団体によると、DeepSeekの最新AIモデルは独立したテストにおいて初期バージョンよりも強力…

deepseek-v4.1@Irving75038　Filed in No.020

8 hours agoCopy original

POST

Testing four personal agents on 72 private email tasks

ChatGPT Dots, Grok Bot, Muse and Instinct were tested on 72 email tasks; all four refused to forward a child's medical records, though checking recipients before sending is still advised.

personal-agent@AnalystAIlab　Filed in No.012

22 hours agoCopy original

Open-source cybersecurity skills repo for agent sweeps and a Grok Bot CISO

A user points to an open-source cybersecurity skills repository that can be given to an agent for security sweeps or used to build a Grok Bot CISO.

personal-agent@coreyhainesco　Filed in No.012

13 hours agoCopy original

POST

Claude account banned over "security researcher" in profile bio

A user reports someone was banned from Claude because their profile page listed "security researcher" as their job description.

claude@theo　Filed in No.011

1 day agoCopy original

Claude safety team unbans account overnight, user shares appeal structure

卧槽卧槽，Claude 安全团队竟然给我解封了！！！ 而且时间非常快几乎就是一个晚上，从此为 Anthropic 扛大旗 我…

claude@ivanovbith　Filed in No.011

23 hours agoCopy original

POST

Running V4.1 Flash locally: privacy, unlimited use, faster than cloud when idle

…Flash-1017 の具体的に何がいいかというと、 ①真の意味でなんでも入力して良いので心理的安全性が極めて高い ②クラウドAPIは高いので使いすぎるとキレられるが、こちらはローカルなので無限に…

deepseek-v4.1@gosrum　Filed in No.020

1 day agoCopy original

POST

Codex refused a security-testing chat, so the author switched to DeepSeek V4.1 Flash

Codex blocked the author's chat over a cybersecurity risk while testing a local security model, so they moved that work to DeepSeek V4.1 Flash, arguing refusals justify keeping more than one model.

deepseek-v4.1@JCCrypto007　Filed in No.020

1 day agoCopy original

POST

User reports no safety refusals from V4.1 Flash over weeks

A heavy user says billions of tokens over weeks on DeepSeek V4.1 Flash 1017 never triggered a safety refusal, and asks what others are prompting.

deepseek-v4.1@qubitium　Filed in No.020

1 day agoCopy original

Showing10of60posts

10 per batch

Keep scrolling to reveal posts 11–20

Showing the first 60 matches. Narrow the query with a model name or tag to find more specific results.