-
Notifications
You must be signed in to change notification settings - Fork 0
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
AISI Inspect integration #6
Conversation
…use. Fixed embedding dimensionality mismatch by specifying embed model name.
…ion is not yet ready.
…xt in solver. Revised nan handling in reducer.
…ics selected in yaml config.
…ue from Inspect plan.
@@ -102,3 +111,70 @@ def get_cfg(self): | |||
if self._cfg is None: | |||
raise ValueError("System config not set.") | |||
return self._cfg.as_dict() | |||
|
|||
def get_inspect_tool( |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
What's the purpose of implementing both the get_inspect_tool
and the get_inspect_solver
. Do we need both, or just one of these is sufficient?
async def solve(state: TaskState, generate: Generate) -> TaskState: | ||
query = state.user_prompt.text | ||
async with concurrency("document_search", max_concurrency): | ||
response = self.invoke(query, documents) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The response
here consists of both the retrieved context and the generated answer, why are we passing it again through the chain_of_thought()
, generate()
and self_critique()
pipeline?
veval/metrics/template.py
Outdated
metric_function=inspect_metric_fn, | ||
)() | ||
for ragas_feature_name in RAGAS_FEATURE_NAMES |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The current implementation calculates all available metrics irrespective of task. Modify this to restrict to only those metrics which are specified in the task config (yaml file).
PR Type
Feature
Short Description
Integrate AISI Inspect for RAG and for RAGAS Evaluation.
To run these examples, install inspect-ai and run: