Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Claude's system prompt is over 24k tokens with tools

github.com

Post reply on HN
Print view View original

31–40 of 350 posts

Re: Claude's system prompt is over 24k tokens with tools

#31

Post by behnamoh »

radeeyate wrote: post #17
eigenblake wrote: post #16

How did they leak it, jailbreak? Was this confirmed? I am checking for the situation where the true instructions are not what is being reported here. The language model could have "hallucinated" its own system prompt instructions, leaving no guarantee that this is the real deal.

All System Prompts from Anthropic models are public information, released by Anthropic themselves: https://docs.anthropic.com/en/release-notes/system-prompts . I'm unsure (I just skimmed through) to what the differences between this and the publicly released ones are, so they're might be some differences.

> The assistant is Claude, created by Anthropic.

> The current date is {{currentDateTime}}.

> Claude enjoys helping humans and sees its role as an intelligent and kind assistant to the people, with depth and wisdom that makes it more than a mere tool.

Why do they refer to Claude in third person? Why not say "You're Claude and you enjoy helping hoomans"?

Re: Claude's system prompt is over 24k tokens with tools

#33

Post by synap5e »

LeoPanthera wrote: post #32

I'm far from an LLM expert but it seems like an awful waste of power to burn through this many tokens with every single request. Can't the state of the model be cached post-prompt somehow? Or baked right into the model?

It's cached. Look up KV (prefix) caching.

Re: Claude's system prompt is over 24k tokens with tools

#34

Post by voxic11 »

LeoPanthera wrote: post #32

I'm far from an LLM expert but it seems like an awful waste of power to burn through this many tokens with every single request. Can't the state of the model be cached post-prompt somehow? Or baked right into the model?

Yes prompt caching is already a widely used technique. https://www.anthropic.com/news/prompt-caching

Re: Claude's system prompt is over 24k tokens with tools

#36

Post by Legend2440 »

4b11b4 wrote: post #10

I like how there are IFs and ELSE IFs but those logical constructs aren't actually explicitly followed... and inside the IF instead of a dash as a bullet point there's an arrow.. that's the _syntax_? hah.. what if there were two lines of instructions, you'd make a new line starting with another arrow..? Did they try some form of it without IFs first?...

Syntax doesn't need to be precise - it's natural language, not formal language. As long as a human could understand it the LLM will too.

Re: Claude's system prompt is over 24k tokens with tools

#37

Post by selectodude »

behnamoh wrote: post #31
radeeyate wrote: post #17

Earlier quoted context omitted.

All System Prompts from Anthropic models are public information, released by Anthropic themselves: https://docs.anthropic.com/en/release-notes/system-prompts . I'm unsure (I just skimmed through) to what the differences between this and the publicly released ones are, so they're might be some differences.

> The assistant is Claude, created by Anthropic. > The current date is {{currentDateTime}}. > Claude enjoys helping humans and sees its role as an intelligent and kind assistant to the people, with depth and wisdom that makes it more than a mere tool. Why do they refer to Claude in third person? Why not say "You're Claude and you enjoy helping hoomans"?

I don’t know but I imagine they’ve tried both and settled on that one.

Re: Claude's system prompt is over 24k tokens with tools

#38

Post by tgtweak »

alchemist1e9 wrote: post #27
SafeDusk wrote: post #14

In addition to having long system prompts, you also need to provide agents with the right composable tools to make it work. I’m having reasonable success with these seven tools: read, write, diff, browse, command, ask, think. There is a minimal template here if anyone finds it useful: https://github.com/aperoc/toolkami

Where does one find the tool prompts that explains to the LLM how to use those seven tools and what each does? I couldn’t find it easily looking through the repo.

You can see it in the cline repo which does prompt based tooling, with Claude and several other models.

Re: Claude's system prompt is over 24k tokens with tools

#39

Post by mplewis »

alchemist1e9 wrote: post #27
SafeDusk wrote: post #14

In addition to having long system prompts, you also need to provide agents with the right composable tools to make it work. I’m having reasonable success with these seven tools: read, write, diff, browse, command, ask, think. There is a minimal template here if anyone finds it useful: https://github.com/aperoc/toolkami

Where does one find the tool prompts that explains to the LLM how to use those seven tools and what each does? I couldn’t find it easily looking through the repo.

You can find these here: https://github.com/search?q=repo%3Aaperoc%2Ftoolkami%20%40mc...

Re: Claude's system prompt is over 24k tokens with tools

#40

Post by llflw »

voxic11 wrote: post #34
LeoPanthera wrote: post #32

I'm far from an LLM expert but it seems like an awful waste of power to burn through this many tokens with every single request. Can't the state of the model be cached post-prompt somehow? Or baked right into the model?

Yes prompt caching is already a widely used technique. https://www.anthropic.com/news/prompt-caching

It seems like it's token caching, not model caching.
Post reply on HN