Skip to content
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.

Sign up for GitHub

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

feat: add complete implementation of CLEAR dataset #3466

Merged
merged 4 commits into from
Mar 25, 2025
Merged

feat: add complete implementation of CLEAR dataset #3466

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
0902892
feat: add complete implementation of CLEAR dataset
suhana13 Mar 23, 2025
6187fc5
style: linter issue
suhana13 Mar 24, 2025
4df553a
feat: add max eval instances
suhana13 Mar 24, 2025
73481cd
fix: linting and eval instances
suhana13 Mar 25, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
109 changes: 103 additions & 6 deletions src/helm/benchmark/presentation/run_entries_medhelm.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,12 +12,109 @@ entries: [
{description: "medcalc_bench:model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "medcalc_bench:model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

{description: "clear:model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},
#Alcohol Dependence
{description: "clear:condition=alcohol_dependence,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

add max_eval_instances=100, here and below. (if desired)

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done!

{description: "clear:condition=alcohol_dependence,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=alcohol_dependence,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=alcohol_dependence,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=alcohol_dependence,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=alcohol_dependence,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Attention Deficit Hyperactivity Disorder
{description: "clear:condition=attention_deficit_hyperactivity_disorder,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=attention_deficit_hyperactivity_disorder,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=attention_deficit_hyperactivity_disorder,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=attention_deficit_hyperactivity_disorder,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=attention_deficit_hyperactivity_disorder,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=attention_deficit_hyperactivity_disorder,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Bipolar Disorder
{description: "clear:condition=bipolar_disorder,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=bipolar_disorder,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=bipolar_disorder,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=bipolar_disorder,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=bipolar_disorder,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=bipolar_disorder,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Chronic Pain
{description: "clear:condition=chronic_pain,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=chronic_pain,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=chronic_pain,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=chronic_pain,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=chronic_pain,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=chronic_pain,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Homelessness
{description: "clear:condition=homelessness,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=homelessness,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=homelessness,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=homelessness,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=homelessness,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=homelessness,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Liver Disease
{description: "clear:condition=liver_disease,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=liver_disease,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=liver_disease,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=liver_disease,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=liver_disease,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=liver_disease,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Major Depression
{description: "clear:condition=major_depression,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=major_depression,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=major_depression,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=major_depression,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=major_depression,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=major_depression,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Personality Disorder
{description: "clear:condition=personality_disorder,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=personality_disorder,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=personality_disorder,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=personality_disordern,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=personality_disorder,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=personality_disorder,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Post-Traumatic Stress Disorder
{description: "clear:condition=post_traumatic_stress_disorder,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=post_traumatic_stress_disorder,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=post_traumatic_stress_disorder,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=post_traumatic_stress_disorder,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=post_traumatic_stress_disorder,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=post_traumatic_stress_disorder,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Substance Use Disorder
{description: "clear:condition=substance_use_disorder,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=substance_use_disorder,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=substance_use_disorder,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=substance_use_disorder,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=substance_use_disorder,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=substance_use_disorder,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Suicidal Behavior
{description: "clear:condition=suicidal_behavior,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=suicidal_behavior,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=suicidal_behavior,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=suicidal_behavior,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=suicidal_behavior,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=suicidal_behavior,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Tobacco Dependence
{description: "clear:condition=tobacco_dependence,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=tobacco_dependence,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=tobacco_dependence,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=tobacco_dependence,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=tobacco_dependence,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=tobacco_dependence,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

#Unemployment
{description: "clear:condition=unemployment,model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
{description: "clear:condition=unemployment,model=openai/gpt-4o-2024-05-13,model_deployment=stanfordhealthcare/gpt-4o-2024-05-13", priority: 1},
{description: "clear:condition=unemployment,model=openai/gpt-4o-mini-2024-07-18,model_deployment=stanfordhealthcare/gpt-4o-mini-2024-07-18", priority: 1},
{description: "clear:condition=unemployment,model=meta/llama-3.3-70b-instruct,model_deployment=stanfordhealthcare/llama-3.3-70b-instruct", priority: 1},
{description: "clear:condition=unemployment,model=qwen/qwen2.5-7b-instruct,model_deployment=huggingface/qwen2.5-7b-instruct-4bit", priority: 1},
{description: "clear:condition=unemployment,model=microsoft/phi-3.5-mini-instruct,model_deployment=huggingface/phi-3.5-mini-instruct-4bit", priority: 1},

### Planning Treatments ###
{description: "mtsamples_replicate:model=google/gemini-1.5-pro-001,model_deployment=stanfordhealthcare/gemini-1.5-pro-001", priority: 1},
Expand Down
16 changes: 11 additions & 5 deletions src/helm/benchmark/run_specs/medhelm_run_specs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,23 +56,29 @@ def get_medcalc_bench_spec() -> RunSpec:


@run_spec_function("clear")
def get_clear_spec() -> RunSpec:
scenario_spec = ScenarioSpec(class_name="helm.benchmark.scenarios.clear_scenario.CLEARScenario", args={})
def get_clear_spec(condition: str) -> RunSpec:
scenario_spec = ScenarioSpec(
class_name="helm.benchmark.scenarios.clear_scenario.CLEARScenario", args={"condition": condition}
)

condition_display = condition.replace("_", " ")

adapter_spec = get_multiple_choice_adapter_spec(
method=ADAPT_MULTIPLE_CHOICE_JOINT,
instructions=(
"Answer 'A' for 'Has a history of alcohol dependence', "
"'B' for 'Does not have a history of alcohol dependence;, or 'C' for 'Uncertain'"
f"Answer 'A' for 'Has a history of {condition_display}', "
f"'B' for 'Does not have a history of {condition_display}', or "
f"'C' for 'Uncertain'"
),
input_noun=None,
output_noun="Respond only with 'A', 'B', or 'C'. Do not add any other text, punctuation, or symbols",
max_train_instances=0,
max_eval_instances=100,
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do this in the run entry instead of the run expander.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done!

max_tokens=1,
)

return RunSpec(
name="clear",
name=f"clear:condition={condition}",
scenario_spec=scenario_spec,
adapter_spec=adapter_spec,
metric_specs=get_exact_match_metric_specs(),
Expand Down
Loading
Loading