Here's more info on how to use it on AWS with Amazon Bedrock:
https://aws.amazon.com/blogs/aws/anthropics-claude-3-5-sonne...
Claude 3.5 Sonnet
131–140 of 287 posts
Re: Claude 3.5 Sonnet
#132>does the period of a point mass moving in a potential well depend on its mass?
"For a simple harmonic oscillator, the period does not depend on the mass of the object. However, for a general potential well, the answer is more nuanced:
Simple harmonic oscillator: In this case, the period T is given by: T = 2π√(m/k) where m is the mass and k is the spring constant. However, the frequency of oscillation ω = √(k/m), so T = 2π/ω. The spring constant k is typically proportional to m, making the period independent of mass.
General potential well: For a non-harmonic potential, the period usually does depend on mass. The period is given by: T = 2∫dx/v(x)
where v(x) is the velocity as a function of position. This velocity depends on the total energy and the potential energy:
v(x) = √(2/m * (E - V(x)))
Here, the mass appears explicitly, and generally doesn't cancel out as it does in the simple harmonic case.
Would you like me to elaborate on any part of this explanation?"
>why do you say the spring constant k is typically proportional to m? a spring's force when stretched to a given length doesn't depend on what's attached to the end
"You're absolutely right, and I apologize for that error in my explanation. You are correct that the spring constant k is a property of the spring itself and does not depend on the mass attached to it."
Re: Claude 3.5 Sonnet
#133It does better on some of my tests but not enough for me to feel confident it's "solving problems". One thing I like to do is pose a problem and give it a bunch of objects it could use, some more helpful than others. I have found language models fail at discerning which of the tools are useful and which are red herrings, opting to use everything in some way. My go to test is the boat and goat: "You are on the shore o…
> Claude 3.5 says to just bring the goat across in the boat, which is wrong but better than previous versions. So that's an improvement. Sometimes it's funny to me how we can have such a feeling the responses are so obviously wrong in some way but then don't even see it the same way between ourselves. Imagine someone strikes up a conversation with you saying they've got a truck & a sofa with them and they want to kno…
For me, the problem with LLMs is their infinite capacity to ad lib and improv; it feels like trying to solve real problems during a taping of "Whose Line is it Anyway?"
Yeah of course it's not a realistic scenario for humans, but the LLM is not a human, it's a tool, and I expect it to have some sort of utility as a tool (repeatability, predictability, fit for purpose). If it can't be used as a tool, and it can't replace human-level inference, then it's worthless at best and antagonistic at worst.
I started testing with the goat/boat prompt because it was obvious given the framing that the LLM was trying to pattern match against the logic problem involving a wolf. Really takes the magic out of it. Most people who hadn't heard the puzzle before would answer with straight up logic, and those who had heard of it would maybe be confused about the framing but wouldn't hallucinate an invisible wolf was part of the solution as so many LLMs do.
To me this just highlights how I have to be an expert at the domain in which I'm prompting, because otherwise I can't be sure the LLM won't suggest I drown a ferret.
Re: Claude 3.5 Sonnet
#134I know it's nitpicking but please anthropic give non business accounts API access.
They don't require it to be a business for API access, they just have an account structure where you can have several users under one 'organization'. Even a family can be an 'organization', nothing special required.
Thanks for the clarification I was sure it meant only business accounts with business tax numbers.
Re: Claude 3.5 Sonnet
#135Opus remained better than GPT for me, even after the release of GPT-4o. VERY happy to see an even further improvement beyond that, Claude is a terrific product and given the news that GPT-5 only began its training several weeks ago I don't see any situation where Anthropic is dethroned in the near term. There are only two parts of Anthropic's offering I'm not a fan of: - Lack of conversation sharing: I had a conversa…
I recently released Slackrock [https://github.com/coreylane/slackrock] that you may find helpful, it's a Slack chat app that can access several FMs (including Claude 3.5) via AWS Bedrock. Responses can be easily shared with others by inviting them to your channels, and Slack has an Android app. It doesn't support attachments (yet) but I'm working on it!
Re: Claude 3.5 Sonnet
#136>does the period of a point mass moving in a potential well depend on its mass? "For a simple harmonic oscillator, the period does not depend on the mass of the object. However, for a general potential well, the answer is more nuanced: Simple harmonic oscillator: In this case, the period T is given by: T = 2π√(m/k) where m is the mass and k is the spring constant. However, the frequency of oscillation ω = √(k/m), so…
Probably a case of combining a first principles explanation with a heuristic explanation for design of a mechanism.
Re: Claude 3.5 Sonnet
#137Opus remained better than GPT for me, even after the release of GPT-4o. VERY happy to see an even further improvement beyond that, Claude is a terrific product and given the news that GPT-5 only began its training several weeks ago I don't see any situation where Anthropic is dethroned in the near term. There are only two parts of Anthropic's offering I'm not a fan of: - Lack of conversation sharing: I had a conversa…
Both GPT-4 and 4o have been completely useless for coding in the past couple of weeks for me - constant errors, and not just your typical LLM inaccuracies but incapable of producing a few lines of self-consistent code e.g. defines variables foo on one line and refers to it as bar on the next, or it misspells it as foox.
Re: Claude 3.5 Sonnet
#138Re: Claude 3.5 Sonnet
#139This is fantastic. I know it's bells-and-whistles and not the core offering, but a thing that keeps me on the $20/month chatgpt (as opposed to using the API) for my use case (mostly interactive data exploration) is its ability to actually run the code it generates. The ergonomics of chatgpt's ability to answer a question by writing and then executing some python and then even displaying a graph are really important to me.
Re: Claude 3.5 Sonnet
#140Earlier quoted context omitted.
From the Anthropic model guide: Agentic Coding Claude 3.5 Sonnet solves 64% of problems on an internal agentic coding evaluation, compared to 38% for Claude 3 Opus. Our evaluation tests a model’s ability to understand an open source codebase and implement a pull request, such as a bug fix or new feature, given a natural language description of the desired improvement. For each problem, the model is evaluated based on…
> The tests are not visible to the model > The model is allowed to write and run code in an agentic loop and iteratively self-correct during evaluation What does this mean? How does the model iteratively self-correct without seeing the tests? Can it see the test results?
It isn’t allowed to see the final evaluation test (used in calculating its pass/fail), but it can run code and see the output of its own code in order to understand what doesn’t work. If it ends up creating tests as part of that based on the original problem statement then presumably that’s allowed.