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

OpenAI's o1 Playing Codenames

suveenellawela.com

Post reply on HN
Print view View original

131–138 of 138 posts

Re: OpenAI's o1 Playing Codenames

#131

Post by hunter2_ »

ta_1138 wrote: post #94
ruds wrote: post #64

Earlier quoted context omitted.

I don't think this sort of communication from guessers to clue giver is in the spirit of the game (at least in my play group). However, inflating later clues is a reasonable approach! It's just that I don't think you're allowed to communicate the amount of inflation. Guessers must determine whether people 5 has slack to allow additional guesses on previous clues.

The communication is only necessary/important if people haven't set this as a convention in the first place. I'll say that prior to ever looking at my clues: "I will give you higher numbers than what I said if you miss by more than 1. THe number I pick will always be high enough as to allow you to, with the +1 guess you get for free, make guesses on all the words I was hinting at. There's also all kinds of not necess…

Indeed, a good Codenames-playing bot should know how to do all of this, in addition to using its LLM to generate great clues.

Re: OpenAI's o1 Playing Codenames

#132

Post by cyode »

lolinder wrote: post #52

A small weakness in this test is that one of the keys to strategic Codenames play is understanding your partner. You're not just trying to connect the words, you're trying to connect them in a way that will be obvious to your partner. As a computing analogy: you're trying to serialize a few cards in a way that will be deserializable by the other player. This test pairs o1 with itself, which means the serializer is th…

Stretching? Never! I see your 4-clue, o1, and raise you “QUEUE” for 5:

  - Line (Standing in the queue…)
  - London (they’re all queued up, innit?)
  - Log (*backend distsys handwaving*)
  - Mail (what do you think an inbox is, anyway?!)
  - Round (homophone “Q” is a typographically round letter)

Re: OpenAI's o1 Playing Codenames

#134

Post by joshvm »

groggo wrote: post #86
jncfhnb wrote: post #58

Earlier quoted context omitted.

If you really want to get good, your goal is not so much to get as many tiles as possible, but rather to get the tiles that are semantically distinct from your opponent’s. A single mistake that triggers your opponent’s tile is generally enough to lose the game. And even if they don’t do it, having them uncover the tiles from their side that are semantically similar to your own team is also useful. If you want to get…

Memorizing the tile layouts is too much for me haha (imo against the spirit of the game). I usually play online now anyway so I hope they don't follow those same patterns as the physical version.

Online specifically avoids this by randomising the grids, except in some modes like mirrors where you can't do much to preserve symmetry.

Re: OpenAI's o1 Playing Codenames

#135

Post by paulddraper »

cyode wrote: post #132
lolinder wrote: post #52

A small weakness in this test is that one of the keys to strategic Codenames play is understanding your partner. You're not just trying to connect the words, you're trying to connect them in a way that will be obvious to your partner. As a computing analogy: you're trying to serialize a few cards in a way that will be deserializable by the other player. This test pairs o1 with itself, which means the serializer is th…

Stretching? Never! I see your 4-clue, o1, and raise you “QUEUE” for 5: - Line (Standing in the queue…) - London (they’re all queued up, innit?) - Log (*backend distsys handwaving*) - Mail (what do you think an inbox is, anyway?!) - Round (homophone “Q” is a typographically round letter)

I think Round may be invalid but in any case I would not have gotten it.

Re: OpenAI's o1 Playing Codenames

#136

Post by jncfhnb »

furyofantares wrote: post #97
jncfhnb wrote: post #57

Earlier quoted context omitted.

Ehhh I don’t think that’s accurate. The problem is not linking 4 words. It’s linking 4 words without accidentally triggering other, semantically adjacent words. This task could probably be solved nearly just as well with old school word 2 vec embeddings

> This task could probably be solved nearly just as well with old school word 2 vec embeddings I've tried. This approach is well beyond awful.

I see a few papers published that did exactly this successfully. It also just sounds crazy that it wouldn’t work well.

It’s odd to me that you would confidently claim it’s “beyond awful”.

Re: OpenAI's o1 Playing Codenames

#137

Post by spencerflem »

hunter2_ wrote: post #128
spencerflem wrote: post #120

Earlier quoted context omitted.

I played in a Codenames tournament at CGE's stand at GenCon, and they forbid guessers from communicating at all. Officially, its supposed to be just the clue and number and nothing else. Of course, I never play this way in my own games

How do guessers arrive at a consensus about what card to touch, if they are forbidden from communicating at all?

officially its a 4 player only game, at least at the tournament. I never do it this way myself though

Re: OpenAI's o1 Playing Codenames

#138

Post by furyofantares »

jncfhnb wrote: post #136
furyofantares wrote: post #97

Earlier quoted context omitted.

> This task could probably be solved nearly just as well with old school word 2 vec embeddings I've tried. This approach is well beyond awful.

I see a few papers published that did exactly this successfully. It also just sounds crazy that it wouldn’t work well. It’s odd to me that you would confidently claim it’s “beyond awful”.

I'm confidently relaying my experience. But I get that I was extremely terse and overly general in my reply.

I haven't surveyed all the papers, although I have read some. And all the ones that I've seen that work okay -- do so by using a language graph or word association graph in their algorithm. Not just embeddings. Even then the results don't look good to me compared to human performance.

Why does it sound crazy that it wouldn't work well? Have you used word embeddings much? Maybe you have and have good reason to think this - I don't mean to imply otherwise. But it doesn't sound crazy to me that it wouldn't work well.

If I am wrong I would love to know it.

Post reply on HN