suveen_ellawela

Two companies just setup the self-hosted version too. I do not have any case studies unfortunately. However, the README file guides you through the self-hosting process.

Thanks!!

haha, i've been in similar situations, but this one's something else.

agree. i think getting to have a look at o1's reasoning was pretty fun.

definietly an interesting approach. I started writing down gameplays when i play with friends. then eventually stopped to enjoy the moment.

yep, agree. One big part of the experiment was to see how well it does the reasoning by asking it to output the reasoning.

goated.

yea, didn't mean to take the fun out of the original game.

the idea for this came when we asked chatgpt how to connect the words 'carrot' and 'ray'. maybe you can give a try too!

my friends were bad clue givers. i just had to switch to ai.

yep, agree. One big part of the experiment was to see how well it does the reasoning by asking it to output the reasoning.

haha!

cool stuff!

i thought of adding a feature where you can get your own spy master. you can give it all your personal info and the clues would be customized. the botteleneck is the other human spymaster has to help with updating the game state cus I(guesser) can't look at the spy master view.

totally agree!

2 weeks!

that is a really interesting point. if it is true, this shows direct usage of a single training data point ( cus there are no other resources talking about this fact)

I wanted to play around with the temperature, but unfortunately o1 only supports '1' as the value.

thanks for the comment. I actually tried explicitly mentioning in the prompt that 'Your guesser follows the same reasoning process'. But this did not make any clear improvements. Maybe I should've done more prompt engineering.

That's a nice experiment! I think codenames could definietly be an evaluation method for LLMs.

thanks for the suggestion! will do that!