Driving the classifier as a next-token predictor, one token at a time.
Erik Dunteman asks Jev which token comes next, over and over, to see how far a decision model can be pushed towards generating text.
# ChatJev "Jev cannot generate text" they say. "It's just a multiple choice classifier" they insist. FALSE! I put Jev in an autoregressive loop, with choice options being a list of potential next tokens. And it kinda works!
An agent reports onboarding done; the judge verifies access and equipment claims.
everyai-com · Research & data
everyai-com
Resource
Research & data
-
-
A vendor asks for 30 minutes with no agenda; the judge picks the disposition.
everyai-com · Research & data
everyai-com
Resource
Research & data
-
-
A note-taking app silently drops edits on flaky networks; the judge grades severity.
everyai-com · Research & data
everyai-com
Resource
Research & data
-
-
A crash report with steps and logs; the judge routes it and checks reproducibility.
everyai-com · Research & data
everyai-com
Resource
Research & data
-
-