[TLM CI] Add more configurable args to unittests

[huiwengoh]

Add args to be able to specify which configs are not expected to produce trustworthiness scores or TLMResponse objects

[github-actions]

Ensure final changes made to the TLM code are tested before merging. To run the TLM tests, comment /test-tlm on this PR.

[huiwengoh]

/test-tlm
✨ Starting TLM tests... ✨
If you want to run the TLM tests again (in case they failed or you have further updates to the TLM code), comment '/test-tlm' on this PR.
View full GitHub Actions run log
✨ Tests completed! ✨
TLM Tests Results: ✅✅ ✅✅✅
TLM Property Tests Results: ❌❌❌❌❌
Click the Github Actions run log for more information.
✨ Tests completed! ✨
TLM Tests Results: ✅✅ ✅✅✅
TLM Property Tests Results: ✅✅✅✅✅
Click the Github Actions run log for more information.
✨ Tests completed! ✨
TLM Tests Results: ❌❌ ❌❌❌
TLM Property Tests Results: ❌❌❌❌❌
Click the Github Actions run log for more information.
✨ Tests completed! ✨
TLM Tests Results: ❌❌ ❌❌❌
TLM Property Tests Results: ❌❌❌❌❌
Click the Github Actions run log for more information.

[huiwengoh]

@ulya-tkch fyi some of the the tests seem to be a little flaky, this first run failed, but when I re-ran it (without changing anything at all), everything passed this time.

My guess is that it might have something to do with the randomization of the args, but I am not sure (I was trying to debug locally but was unable to reproduce as well). Just flagging to keep an eye out in the future.

[huiwengoh]

✨ Tests completed! ✨
TLM Tests Results: ❌❌ ❌❌❌
TLM Property Tests Results: ❌❌❌❌❌
Click the Github Actions run log for more information.

I am also unsure what happened here? Is it because I added a comment?

Edit: I think it is bc I added a comment, a second set of that has appeared haha

[ulya-tkch]

Yeah so the PR is merged to prevent that from happening, but the results were posted everytime a comment was added. This should no longer happen. Curent comments:

✨ Tests completed! ✨
TLM Tests Results: ✅✅ ✅✅✅
TLM Property Tests Results: ❌❌❌❌❌
Click the Github Actions run log for more information.
✨ Tests completed! ✨
TLM Tests Results: ✅✅ ✅✅✅
TLM Property Tests Results: ✅✅✅✅✅
Click the Github Actions run log for more information.
✨ Tests completed! ✨
TLM Tests Results: ❌❌ ❌❌❌
TLM Property Tests Results: ❌❌❌❌❌
Click the Github Actions run log for more information.
✨ Tests completed! ✨
TLM Tests Results: ❌❌ ❌❌❌
TLM Property Tests Results: ❌❌❌❌❌
Click the Github Actions run log for more information.

[ulya-tkch]

[nit] Shouldn't _test_batch_get_trustworthiness_score_response(responses, options) also be updated for models with no perplexity score?

[huiwengoh]

FYI, I am not changing the is_trustworthiness_score functions bc I think #304 is introducing much larger changes