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

LLMs get lost in multi-turn conversation

arxiv.org

Post reply on HN
Print view View original

211–220 of 272 posts

Re: LLMs get lost in multi-turn conversation

#211

Post by sixtyj »

johnisgood wrote: post #160
sixtyj wrote: post #133

Earlier quoted context omitted.

In Gemini, it is really good to have large window with 1M tokens. However, around 100,000 it starts to make mistakes and refactor its own code. Sometimes it is good to start new chat or switch to Claude. And it really helps to be very precise with wording of specification what you want to achieve. Or repeat it sometimes with some added request lines. GIGO in reality :)

Oh my, I hate it when it rewrites >1k LOC. I have to instruct it to "modify only ..., do not touch the rest" and so forth, but GPT does not listen to this often, Claude does. I dunno about Gemini.

I received this gem in Gemini right now:

I am giving up on providing code, and on checking is it working, because it is very time consuming. Tell me when it starts working. Good luck.

:)

Re: LLMs get lost in multi-turn conversation

#212

Post by Retric »

wrasee wrote: post #148
skydhash wrote: post #51

Earlier quoted context omitted.

Or you could just read the PPP RFC [0]. I’m not saying that your approach is wrong. But most LLM workflows are either brute forcing the solution, or seeking a local minima to be stuck in. It’s like doing thousands of experiments of objects falling to figure out gravity while there’s a physics textbooks nearby. [0]: https://datatracker.ietf.org/doc/html/rfc1661

I could also walk everywhere, but sometimes technology can help. There’s no way I could fully read that RFC in an hour. And that’s before you even know what reading to focus your attention on, so you’re just being a worse LLM at that point.

The difference is you’d remember some of the context from reading the thing where an LLM is starting from scratch every single time it comes up.

Re: LLMs get lost in multi-turn conversation

#213

Post by maleldil »

tmountain wrote: post #159

I often ask the LLM for a concise summary of the discussion so far—formatted as a prompt. I then edit it appropriately and use it to start a new conversation without the baggage. I have found this to be a very effective technique, but I imagine it will be automated sometime soon.

Claude Code has a /compact command that summarises the conversation so far to save on context tokens.

Re: LLMs get lost in multi-turn conversation

#214

Post by Terr_ »

famouswaffles wrote: post #147
littlestymaar wrote: post #115

Earlier quoted context omitted.

It's not a tired argument, and not just a semantic one it's a foundational characteristic of LLM. > A token-predictor could still be trained to predict the tokens “I’m not sure what you mean because of points x, y, and z; could you elaborate?” This is entirely true, and the key insight is even right in your sentence but you don't seem to grasp it. “could still be trained”: you can train an LLM into doing whatever you…

It's definitely a tired and semantical one because as he said, it brings no insight and is not even good at the analogy level. I can't have a conversation with Dracula and Dracula can't make decisions that affect the real world, so LLMs already break key aspects and assumptions of the 'Document Simulator'. Pre-trained LLMs will ask clarifying questions just fine. So I think this is just another consequence of post-tr…

> Dracula can't make decisions that affect the real world, so LLMs already break key aspects and assumptions of the 'Document Simulator'.

Nonsense, we are already surrounded by mindless algorithms (and their outputs) that "affect the real world" because many of us have full-time jobs ensuring it happens! "

When someone uses a SimCity-esque program to generate a spreadsheet used for real-world bus schedules, does that "break key aspects and assumptions of a traffic simulator"? Does the downstream effect elevate it to a microcosm of tiny lives? Nope!

Re: LLMs get lost in multi-turn conversation

#215

Post by johnisgood »

sixtyj wrote: post #211
johnisgood wrote: post #160

Earlier quoted context omitted.

Oh my, I hate it when it rewrites >1k LOC. I have to instruct it to "modify only ..., do not touch the rest" and so forth, but GPT does not listen to this often, Claude does. I dunno about Gemini.

I received this gem in Gemini right now: I am giving up on providing code, and on checking is it working, because it is very time consuming. Tell me when it starts working. Good luck. :)

It is right, it is time consuming. I do not blame it. :D

Re: LLMs get lost in multi-turn conversation

#216

Post by tough »

Helmut10001 wrote: post #14
Benjammer wrote: post #2

It's nice to see a paper that confirms what anyone who has practiced using LLM tools already knows very well, heuristically. Keeping your context clean matters, "conversations" are only a construct of product interfaces, they hurt the quality of responses from the LLM itself, and once your context is "poisoned" it will not recover, you need to start fresh with a new chat.

My experiences somewhat confirm these observations, but I also had one that was different. Two weeks of debugging IPSEC issues with Gemini. Initially, I imported all the IPSEC documentation from OPNsense and pfSense into Gemini and informed it of the general context in which I was operating (in reference to 'keeping your context clean'). Then I added my initial settings for both sides (sensitive information redacted!…

LLM's are good at interpolation but bad at extrapolating

Re: LLMs get lost in multi-turn conversation

#217

Post by djmips »

TheOtherHobbes wrote: post #123
kfarr wrote: post #109

Earlier quoted context omitted.

Who gets lost in multi-turn conversations?

Everyone? How often in meetings does everyone maintain a running context of the entire conversation, instead of responding to the last thing that was said with a comment that has an outstanding chance of being forgotten as soon as the next person starts speaking?

Indeed - and since human's are susceptible to injection prompt, all it needs is one derailing comment to take things off course.

Re: LLMs get lost in multi-turn conversation

#218

Post by cgriswald »

Retric wrote: post #212
wrasee wrote: post #148

Earlier quoted context omitted.

I could also walk everywhere, but sometimes technology can help. There’s no way I could fully read that RFC in an hour. And that’s before you even know what reading to focus your attention on, so you’re just being a worse LLM at that point.

The difference is you’d remember some of the context from reading the thing where an LLM is starting from scratch every single time it comes up.

There are opportunity costs to consider along with relevance. Suppose you are staying at my place. Are you going to read the manual for my espresso machine in total or are you going to ask me to show you how to use it or make one for you?

In any case, LLMs are not magical forgetfulness machines.

You can use a calculator to avoid learning arithmetic but using a calculator doesn’t necessitate failing to learn arithmetic.

You can ask a question of a professor or fellow student, but failing to read the textbook to answer that question doesn’t necessitate failing to develop a mental model or incorporate the answer into an existing one.

You can ask an LLM a question and blindly use its answer but using an LLM doesn’t necessitate failing to learn.

Re: LLMs get lost in multi-turn conversation

#219

Post by daveguy »

tough wrote: post #216
Helmut10001 wrote: post #14

Earlier quoted context omitted.

My experiences somewhat confirm these observations, but I also had one that was different. Two weeks of debugging IPSEC issues with Gemini. Initially, I imported all the IPSEC documentation from OPNsense and pfSense into Gemini and informed it of the general context in which I was operating (in reference to 'keeping your context clean'). Then I added my initial settings for both sides (sensitive information redacted!…

LLM's are good at interpolation but bad at extrapolating

To be fair, all AI/ML and even statistical methods are bad at extrapolating.

Re: LLMs get lost in multi-turn conversation

#220

Post by tough »

sixtyj wrote: post #211
johnisgood wrote: post #160

Earlier quoted context omitted.

Oh my, I hate it when it rewrites >1k LOC. I have to instruct it to "modify only ..., do not touch the rest" and so forth, but GPT does not listen to this often, Claude does. I dunno about Gemini.

I received this gem in Gemini right now: I am giving up on providing code, and on checking is it working, because it is very time consuming. Tell me when it starts working. Good luck. :)

I love it when models give up, gives me some hope humans will still be required for the time being lol
Post reply on HN