Teaching Alpacas to Act like a classifier

preamble

After a little bit of experimentation, I think I have found a way to get Stanford's Alpaca 7B model to give me a reasonable result.

Input

Below is an instruction that describes a task. Write a response that appropriately completes the request. ### Instruction: the following are the changes for the latest release:

<paste all the changelog entries here>
 - Provide a component stack as a second argument to onRecoverableError. = [Added] - Fix hydrating into document causing a blank page on mismatch. = [Fixed] - Fix false positive hydration errors with Suspense. = [Fixed] - Fix ignored setState in Safari when adding an iframe. = [Fixed]. Should this release be classified as Major, Minor, or Patch?
<changelog entries>

 ### Response: this release should be classified as

Results

The autocompletion gave me:

Below is an instruction that describes a task. Write a response that appropriately completes the request. ### Instruction: the following are the changes for the latest release: - Provide a component stack as a second argument to onRecoverableError. = [Added] - Fix hydrating into document causing a blank page on mismatch. = [Fixed] - Fix false positive hydration errors with Suspense. = [Fixed] - Fix ignored setState in Safari when adding an iframe. = [Fixed]. Should this release be classified as Major, Minor, or Patch? ### Response: this release should be classified as �Minor Release��

the phrase "Minor Release" is actualy rather effective, as we can just grep for the keywords we need and it effecively becomes a rather unhinged black box classifier.

why does it work?