Feature description
We should be able to set the effort parameter as part of the wags llm task runner.
https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-anthropic-claude-messages-request-response.html#model-parameters-anthropic-claude-effort-parameter
This can be set to high, medium, or low and controls token usage for thinking and checks on Claude's side.
Use case
We want to test how effort will affect classification accuracy for prognostic evidence
Acceptance Criteria
Given I am trying to run a classification query via wags-llm in an analysis notebook
When I define my parameters
Then I can set effort=high/medium/low to adjust my token usage
And see how it affects my classification accuracy
Proposed solution
maybe just a EFFORT=(high/medium/low) as part of the other parameters from the task runner
Potential Impact
No response
Feature description
We should be able to set the effort parameter as part of the wags llm task runner.
https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-anthropic-claude-messages-request-response.html#model-parameters-anthropic-claude-effort-parameter
This can be set to high, medium, or low and controls token usage for thinking and checks on Claude's side.
Use case
We want to test how effort will affect classification accuracy for prognostic evidence
Acceptance Criteria
Given I am trying to run a classification query via wags-llm in an analysis notebook
When I define my parameters
Then I can set effort=high/medium/low to adjust my token usage
And see how it affects my classification accuracy
Proposed solution
maybe just a EFFORT=(high/medium/low) as part of the other parameters from the task runner
Potential Impact
No response