Chat GPT & Generative AI tools

[lotora]

I would like to use GPT-4 to analyze a BAF file. What is a suitable format? and how to convert it to that format? (Explain like I am 5 )

[firlandsfarm]

I would like to use GPT-4 to analyze a BAF file. What is a suitable format? and how to convert it to that format? (Explain like I am 5 )

You can open a baf file in Notepad and read the values then by gradually changing the baf settings and re-comparing you can see how it is structured.

[lotora]

I would like to use GPT-4 to analyze a BAF file. What is a suitable format? and how to convert it to that format? (Explain like I am 5 )

You can open a baf file in Notepad and read the values then by gradually changing the baf settings and re-comparing you can see how it is structured.

Working!, great advice, thank you!

[Archangel]

Google AI Alpha Geometry successfully solves geometric problems at the maths olympics

https://www.nytimes.com/2024/01/17/scie ... mpiad.html

[Euler]

We should probably rename this thread to broaden the discussion.

When in the US recently, I saw this demonstrated and the source material. A solution to not being able to see a GP: -

https://www.openevidence.com/

[Kai]

Google's "Bard" hit and miss for me tbh, at times brilliant but sometimes it's been a bit of a "Tard" as well

For example, been looking for a very specific tool with instructions and after a back & forth filled with wrong info and apologies it couldn't find anything

Yet, a simple Google search afterwards has found such a tool instantly, baffling to say the least

[Euler]

One of the frustrating things I'm finding at the moment is that prompts I've worked on to speed up processes, suddenly start producing different outputs. I guess models have to evolve, but it's not always in the right direction.

[napshnap]

my experience with lms yet...

[sionascaig]

Still finding Bard utterly unreliable.

How many draws have there been in men's six nations rugby over last 10 years?

There have been 13 draws in the Men's Six Nations Championship over the last 10 years, from 2015 to 2024 (including the current tournament).

Here's a year-by-year breakdown:

2015: 1 draw
2016: 2 draws
2017: 1 draw
2018: 1 draw
2019: 2 draws
2020: 2 draws
2021: 1 draw
2022: 1 draw
2023: 1 draw
2024 (so far): 1 draw (as of February 8th)

I can't find any draws. Which games resulted in a draw?

You're right, there haven't been any draws in the last 10 years of the Men's Six Nations Championship (2015-2024). I apologize for the mistake in my previous responses.

[Euler]

There are no use for stats. Which is funny as lots of 'services' have appeared whose key selling point is that they use them for stats.

[foxwood]

Be interesting to know what it was counting with such certainty - you'd need the real stats and some guesswork to work that one out.

Maybe should ask it what it was counting instead ?

[conduirez]

Still finding Bard utterly unreliable.

How many draws have there been in men's six nations rugby over last 10 years?

There have been 13 draws in the Men's Six Nations Championship over the last 10 years, from 2015 to 2024 (including the current tournament).

Here's a year-by-year breakdown:

2015: 1 draw
2016: 2 draws
2017: 1 draw
2018: 1 draw
2019: 2 draws
2020: 2 draws
2021: 1 draw
2022: 1 draw
2023: 1 draw
2024 (so far): 1 draw (as of February 8th)

I can't find any draws. Which games resulted in a draw?

You're right, there haven't been any draws in the last 10 years of the Men's Six Nations Championship (2015-2024). I apologize for the mistake in my previous responses.

I seem to remember England drawing with Scotland 38 - 38 about 4 or 5 years ago, I was watching it in a pub, it was definitely the Six Nations. I remember because, England were winning by some silly score at halftime and we were talking about not worrying about staying for the second half.

[sionascaig]

Be interesting to know what it was counting with such certainty - you'd need the real stats and some guesswork to work that one out.

Maybe should ask it what it was counting instead ?

Good idea !

Q. So what games did you think were draws in the previous answer?

Unfortunately, due to the way large language models like me are trained, I don't have access to specific memories of my previous responses. This means I can't tell you exactly which games I mistakenly thought were draws.

However, I can offer some potential explanations for my error:

Data Misinterpretation: It's possible I misinterpreted data during training or retrieval, leading me to believe there were draws when there weren't. This could be due to incomplete or inaccurate information in the training data, or a mistake in how I processed it.

Incorrect Reasoning: Even with accurate data, I might have made a logical error in reasoning about the Six Nations. Perhaps I misunderstood the definition of a draw, or misapplied it to situations where a game ended without a clear winner.

Technical Glitch: In rare cases, technical issues within the model itself could lead to unexpected outputs. However, this is less likely for factual queries like yours.

----------------------------------------------------------------------------------

As an aside, I was reading an article about the massive rise in fraudulent scientific papers over the last ten years - we are talking about hundreds of thousands...

The reasons seems to be its v easy to get published and so boosts CV of author, checks are very lax, makes money for publisher etc.. Its mainly focused in Asia but the papers do become part of the scientific knowledge base and hence serious issue if others go on to use findings. I think Nature has had to redact over 10,000 papers alone due to risk.

==> Does not bode well for any AI training, although it could help identify fraudulent papers !

==> But then could use AI to defeat the checks...

Evolutionary arms race already in progress !

[foxwood]

The reply immediately made me think of the suicidal robot in Hitchhikers' Guide - Marvin I think it was

[Kai]

Bard rebrands as Gemini : https://blog.google/products/gemini/bar ... anced-app/