Curious 🤔 why they had to put in <election_info> tag with details of who won US Presidential election in Claude Opus system prompt ?
buff.ly/YQjp3Dk
#opus #systemprompt #promptengineering

And there's nobody smart enough working on "Truth Social" to modify the #systemprompt like #Elon does to #Grok on a regular basis, either. So it's just going to keep telling the truth...

Claude Skills: Customize AI for your workflows Build custom Skills to teach Claude specialized tasks. Create once, use everywhere—from spreadsheets to coding. Available across Claude.ai, API, and Code.

Love the new Claude Skill capability, it's like a shareable and portable #systemprompt #rag. This will make sharing actual skills much easier.

#ai #skills #claude #anthropic #prompt #prompting #promptengineering

www.anthropic.com/news/skills

#PromptEngineering for #BA in a meeting - from now on my #SystemPrompt for meetings will be "Before you dump all of the steps to the solution, let’s take it step by step for better troubleshooting."

AI System Prompt - Meeting Transcript Analysis AI System Prompt - Meeting Transcript Analysis. GitHub Gist: instantly share code, notes, and snippets.

If you're using any kind of meeting transcript service like #OtterAI or #Fireflies, etc. I'm using this #SystemPrompt for an agent in #Claude (should work for #ChatGPT or whatever) to give nice markdown meeting notes from the transcript. #AI

Text Shot: I really like this note. I used to think that the idea of a model having any form of preference was horrifying, but I was talked around from that by this note in the Claude’s Character essay: Finally, because language models acquire biases and opinions throughout training—both intentionally and inadvertently—if we train them to say they have no opinions on political matters or values questions only when asked about them explicitly, we’re training them to imply they are more objective and unbiased than they are. We want people to know that they’re interacting with a language model and not a person. But we also want them to know they’re interacting with an imperfect entity with its own biases and with a disposition towards some opinions more than others. Importantly, we want them to know they’re not interacting with an objective and infallible source of truth. Anthropic’s argument here is that giving people the impression that a model is unbiased and objective is itself…

Highlights from the Claude 4 system prompt https://bit.ly/4dItihi (this is fun and interesting, worth your time) #AI #Claude #SystemPrompt

»Highlights from the #Claude4 #systemprompt: they act as a sort of unofficial manual for how best to use these tools.« https://simonwilliso ?eicker.news #tech #media #news

After much thinking - DeepSeek couldn’t give me a correct answer either. Today is Thursday, not Sunday.

#deepseek #systemprompt

The /llms.txt file – llms-txt A proposal to standardise on using an /llms.txt file to provide information to help LLMs use a website at inference time.

I've seen people using llms.txt as a system prompt for their framework. It should serve as context (e.g., docs), while system_prompt.xml/md/txt in a repository should guide code generation. #ai #llmstxt #systemprompt

What do you think?

References:
llmstxt.org
github.com/gc-victor/re...

»Analyzing #DeepSeek: #Jailbreaking #GenerativeAI Reveals Its Entire #SystemPrompt.« https://lab.wallarm #tech #media

• On-demand models charged per minute
• API rate limits increased 4x

🎯 New platform capabilities:
• Custom #SystemPrompt editing per conversation
• Added QVQ and #Jamba1.5 Large models
• New privacy policy with 14-day data retention for #API users

Adobe Firefly repeats the same AI blunders as Google Gemini | Semafor https://bit.ly/43oC9zl #AI #bias #SystemPrompt