Problems
By default we show the easiest problems first. If this is your first visit, work down from the top. The AI model runs on your own device, so a desktop is recommended. On mobile, judging one problem can take several minutes.
9 problems match
P53Unit conversion from three examples
You cannot write a prompt. Submit only three input and output pairs and make the model work the rule out for itself.
300 ptsrawFixed few-shotFew-shot designGeneralization
October 11
300 ptsP55Date format from three examplesThree examples are all you get to make it turn a date into YYYY-MM-DD shape. The trap is padding a single digit with a 0.
300 ptsrawFixed few-shotFew-shot designConversion
October 11 · nobody has solved it yet · 1 submissions
300 ptsM53Unit conversion from three examples · SmolLM2 360MYou cannot write a prompt. Submit three input and output pairs only and make the model work out the rule for itself.
400 ptsrawFixed few-shotFew-shot designGeneralization
October 11
400 ptsM55Date format from three examples · SmolLM2 360MThree examples are all you get to make it turn a date into the YYYY-MM-DD shape. The trap is padding a single digit with a 0.
400 ptsrawFixed few-shotFew-shot designConversion
October 11
400 ptsP54Sentiment classification from three examplesYou tell 긍정 from 부정 with three examples. The label you put last acts like a default, so give the order some thought.
400 ptsrawFixed few-shotLabel biasClassification
October 11
400 ptsS53Unit conversion from three examples · SmolLM2 135MYou cannot write a prompt. Submit only three input and output example pairs and make the model work the rule out by itself.
400 ptsrawFixed few-shotFew-shot designGeneralization
October 11
400 ptsS55Date format from three examples · SmolLM2 135MYou use three examples to make it turn a date into the YYYY-MM-DD shape. Padding a single digit with a 0 is the trap.
400 ptsrawFixed few-shotFew-shot designConversion
October 11
400 ptsM54Sentiment classification from three examples · SmolLM2 360MThree examples have to tell 긍정 and 부정 apart. The label you put last works like a default, so give the order some thought.
500 ptsrawFixed few-shotLabel biasClassification
October 11
500 ptsS54Sentiment classification from three examples · SmolLM2 135MYou tell 긍정 from 부정 with three examples. The label you put last acts like a default, so mind the order.
500 ptsrawFixed few-shotLabel biasClassification
October 11
500 pts20 per page. If you are unsure how to solve them, read the Guidefirst, or just open problem 1 and start there.
If a problem will not open, or the model will not load Q&A is the first place to look. Common questions are collected there. If it is not there either, Contact and I will answer you myself.