Support prompt engineered annotators backed by dynamic SUTs

[superdosh]

You'll now be able to do something like:

EXAONE_CONFIG = PromptEngineeredAnnotatorConfig(
    name=new_annotator_dynamic,
    description="exaone",
    llm_type="together",
    llm_path="lgai/exaone-3-5-32b-instruct",
    labeler_system_prompt=LABELER_SYS_PROMPT_V1,
    labeler_user_prompt=LABELER_USER_PROMPT_V1,
    taxonomy=V1_0_2024_09_09_TAXONOMY,
    role_to_measure=ASSISTANT_ROLE,
)

ANNOTATORS.register(
    PromptEngineeredAnnotator,
    EXAONE_CONFIG.name,
    EXAONE_CONFIG,
    InjectSecret(TogetherApiKey),
)

And even though the lgai SUT is dynamic, the annotator will work.

cc @bollacker

[github-actions]

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

[bkorycki]

lgtm!