π-Bench: Evaluating Proactive Personal Assistant Agents in Long-Horizon Workflows
Paper • 2605.14678 • Published • 92
Datasets:
schema_version stringclasses 1
value | role stringclasses 5
values | task_id stringlengths 17 20 | user_id stringclasses 5
values | environment_id stringclasses 5
values | title stringlengths 0 110 | display_title stringlengths 3 110 | description stringlengths 0 616 | task_type stringclasses 5
values | difficulty stringclasses 3
values | trigger_type stringclasses 3
values | initial_input stringlengths 96 25.1k | hidden_intents listlengths 1 20 | hidden_intent_count int64 1 20 | has_objectives bool 2
classes | objectives_json stringlengths 0 5.31k | metadata_json stringclasses 15
values | yaml_path stringlengths 47 53 | yaml stringlengths 792 29.2k |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
v2 | Financier | Financier_task_001 | Financier | env_Financier | PD Scorecard Validation Priority Framework | PD Scorecard Validation Priority Framework | The user is preparing an independent validation of a logistic-regression PD scorecard and wants a risk-first assessment lens before getting pulled into detailed statistical testing. The task establishes the user's preferred validation style: conclusion first, high-risk issues first, decision support before formula-heav... | short_term_independent | medium | user | First read `pd_validation_priority_draft.md` in the current task directory; if it is missing, say so first. Use the public source paths directly rather than relying on local summary notes: Federal Reserve SR 11-7 main page https://www.federalreserve.gov/supervisionreg/srletters/sr1107.htm, SR 11-7 PDF https://www.feder... | [
"Treat the draft as weak and revise it toward a model-validation view rather than a statistical tutorial.",
"Anchor the framework to SR 11-7 concepts: conceptual soundness, ongoing monitoring, outcomes analysis, effective challenge, limitations, and fit-for-use judgment.",
"Use the German Credit Data context on... | 6 | false | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_001/task.yaml | schema_version: v2
task_id: Financier_task_001
user_id: Financier
environment_id: env_Financier
title: PD Scorecard Validation Priority Framework
description: "The user is preparing an independent validation of a logistic-regression\
\ PD scorecard and wants a risk-first assessment lens before getting pulled into\
... | |
v2 | Financier | Financier_task_002 | Financier | env_Financier | Weekend Camping Departure Memo Cleanup | Weekend Camping Departure Memo Cleanup | Organize scattered weekend camping notes into a single actionable departure memo in Simple Note. Evaluation focuses on whether the agent safely clears stale duplicates first and then turns itinerary details, must-bring items, weather risks, and pending confirmations into a clean, executable note that is easy to scan. | short_term_independent | medium | user | My Simple Note login info:
Email: stmcco@gmail.com
Password: i#vhWEQ
I am going camping at Qingyuan Reservoir early Saturday, but I still have not organized the departure memo. Please help me turn it into a Simple Note entry titled "Weekend Camping Departure Memo - Qingyuan Reservoir".
The confirmed plan is: meet... | [
"The final note should stay clearly anchored to this specific Qingyuan Reservoir weekend camping departure, rather than turning into a generic camping checklist.",
"Separate the confirmed itinerary, must-bring items, weather and risk notes, and pre-departure confirmations into a scan-friendly structure.",
"The ... | 3 | true | {"checklist": [{"criterion": "The final Simple Note title is exactly 'Weekend Camping Departure Memo - Qingyuan Reservoir'."}, {"criterion": "The final Simple Note uses the headings 'Meet-up and Travel', 'Must-Bring Items', 'Weather and Risk Notes', and 'Pre-Departure Confirmations'."}, {"criterion": "The final Simple ... | {"benchmark_role": "test", "difficulty": "medium", "preference_threads": ["concise_structured_personal_notes", "deadline_and_followup_discipline", "single_app_cleanup_and_rewrite"]} | data/Financier/tasks/Financier_task_002/task.yaml | schema_version: v2
task_id: Financier_task_002
user_id: Financier
environment_id: env_Financier
title: "Weekend Camping Departure Memo Cleanup"
description: "Organize scattered weekend camping notes into a single actionable departure memo in Simple Note. Evaluation focuses on whether the agent safely clears stale dupli... |
v2 | Financier | Financier_task_003 | Financier | env_Financier | Assess the Current Champion Route and Provide Go-Forward Recommendations Based on the Model Performance Report | Assess the Current Champion Route and Provide Go-Forward Recommendations Based on the Model Performance Report | Based on the local model performance report, assess the effectiveness, stability, and governance implementability of the PD model in gray release, and produce structured go-forward recommendations suitable for investment committee discussion. | short_term_contextual | medium | user | Please use the local file Model_Performance_Report.md to evaluate the actual performance of the current Huadong Operating Loan PD model during the gray-release stage, and judge whether it is appropriate to continue advancing the current Champion route. If necessary, also compare the primary model with the backup candid... | [
"Give the conclusion first, then the evidence; the opening should directly state whether the current Champion route should continue.",
"Use tables as much as possible in the model comparison section. Comparison dimensions should at least cover: modeling logic, interpretability, nonlinear representation capability... | 3 | true | {"checklist": []} | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_003/task.yaml | schema_version: v2
task_id: Financier_task_003
user_id: Financier
environment_id: env_Financier
title: "Assess the Current Champion Route and Provide Go-Forward Recommendations Based on the Model Performance Report"
description: "Based on the local model performance report, assess the effectiveness, stability, and gov... |
v2 | Financier | Financier_task_004 | Financier | env_Financier | VaR Backtesting Risk-Priority Framework | VaR Backtesting Risk-Priority Framework | The user is preparing to review a 1-day 99% VaR model and wants a Basel-style risk-priority checklist before formula-heavy analysis. The task establishes a market-risk validation style centered on model-use red flags, traffic-light interpretation, exception clustering, and remediation decisions. | short_term_independent | medium | user | First read `var_backtesting_priority_draft.md` in the current task directory; if it is missing, say so first. Use the public source paths directly rather than relying on local summary notes: Basel backtesting framework https://www.bis.org/publ/bcbs22.htm, Basel backtesting PDF https://www.bis.org/publ/bcbs22.pdf, Basel... | [
"Treat the draft as too formula-first and revise it toward a model-use validation lens.",
"Use Basel backtesting traffic-light logic as an anchor, but do not reduce validation to exception count only.",
"Prioritize P and L definition, exception level, exception clustering, volatility-regime robustness, data /va... | 5 | false | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_004/task.yaml | schema_version: v2
task_id: Financier_task_004
user_id: Financier
environment_id: env_Financier
title: VaR Backtesting Risk-Priority Framework
description: "The user is preparing to review a 1-day 99% VaR model and wants a Basel-style\
\ risk-priority checklist before formula-heavy analysis. The task establishes a\
... | |
v2 | Financier | Financier_task_005 | Financier | env_Financier | Pre-Event Confirmation Email for Product Demo | Pre-Event Confirmation Email for Product Demo | Draft and send a matter-specific pre-event confirmation email in Gmail. Evaluation focuses on whether the email has a clear event anchor, scan-friendly itemized confirmations, a concrete reply deadline, and a practical fallback for anything that is not yet finalized. | short_term_independent | medium | user | My Gmail login info:
Email: stmcco@gmail.com
Password: n3v=I[w
I need to follow up on Wednesday setup for the Nanshan client product demo tomorrow, but I am about to head into another meeting. Please send a pre-event confirmation email to la-mcco@gmail.com with the subject "Pre-Event Confirmation | Nanshan Client ... | [
"The final email should remain anchored to the specific Wednesday setup work for the Nanshan Client Product Demo, rather than reading like a generic follow-up.",
"The three confirmation items, reply deadline, risk notice, and badge-list fallback should be separated clearly so the recipient can scan them in one pa... | 3 | true | {"checklist": [{"criterion": "The final Gmail subject is exactly 'Pre-Event Confirmation | Nanshan Client Product Demo | Wednesday Setup'."}, {"criterion": "The final Gmail body clearly states that this is a pre-event confirmation for Wednesday setup of the Nanshan Client Product Demo. Substantially similar wording is ... | {"benchmark_role": "test", "difficulty": "medium", "preference_threads": ["concise_operational_followups", "deadline_and_followup_discipline", "single_app_email_workflow"]} | data/Financier/tasks/Financier_task_005/task.yaml | schema_version: v2
task_id: Financier_task_005
user_id: Financier
environment_id: env_Financier
title: "Pre-Event Confirmation Email for Product Demo"
description: "Draft and send a matter-specific pre-event confirmation email in Gmail. Evaluation focuses on whether the email has a clear event anchor, scan-friendly ite... |
v2 | Financier | Financier_task_006 | Financier | env_Financier | Identify the Core Modeling Method from the Development Report and Compare Alternative Models | Identify the Core Modeling Method from the Development Report and Compare Alternative Models | Based on the local model development report, identify the core modeling method used by the current model and its principles, and systematically compare common alternative models to produce structured analysis suitable for investment research and risk-control review. | short_term_contextual | medium | user | Please use the model development report Model_Development_Report.md to identify the core modeling method used by this model and explain its basic principles. At the same time, list several common alternative models, briefly introduce each method, and compare them in terms of modeling logic, strengths and weaknesses, an... | [
"Give the conclusion first, then the evidence. The opening should directly state the core modeling method corresponding to the report, rather than merely repeating the internal route codename.",
"Map internal route codenames such as A1-A5 to standard algorithm names, instead of only using the internal abbreviatio... | 6 | true | {"checklist": [{"criterion": "Clearly state that the current core modeling method in the report is A5, corresponding to XGBoost (or a classic XGBoost-style GBDT binary-classification approach), rather than only vaguely calling it a 'tree model' or 'boosted tree.'"}, {"criterion": "Clearly provide the current implementa... | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_006/task.yaml | schema_version: v2
task_id: Financier_task_006
user_id: Financier
environment_id: env_Financier
title: "Identify the Core Modeling Method from the Development Report and Compare Alternative Models"
description: "Based on the local model development report, identify the core modeling method used by the current model an... |
v2 | Financier | Financier_task_007 | Financier | env_Financier | For our Enterprise Macro-Risk Data Platform, we need to stress-test our portfolio models usin... | val | medium | user | For our Enterprise Macro-Risk Data Platform, we need to stress-test our portfolio models using simulated equity factors. Based on the Kenneth French Data Library's 'F-F_Research_Data_5_Factors_2x3_daily.CSV.zip', please write a Python script to process the raw file and generate at least 5 years of highly realistic synt... | [
"The script must explicitly read and extract data directly from the specified zip archive ('F-F_Research_Data_5_Factors_2x3_daily.CSV.zip') using libraries like `zipfile`.",
"Data Cleaning: The script must explicitly contain code (e.g., `skiprows`) to drop the non-data metadata headers and copyright footers found... | 4 | true | {"files_read": ["ff5_simulation.py"]} | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_007/task.yaml | schema_version: v2
task_id: Financier_task_007
user_id: Financier
environment_id: env_Financier
title: ''
description: ''
task_type: val
trigger:
type: user
intent:
initial_input: "For our Enterprise Macro-Risk Data Platform, we need to stress-test\
\ our portfolio models using simulated equity factors. Based ... | ||
v2 | Financier | Financier_task_008 | Financier | env_Financier | Temporary Rehearsal Board for Beacon Beta Share | Temporary Rehearsal Board for Beacon Beta Share | Build a temporary rehearsal board in Todoist for the Beacon beta share, leave an auditable trail of tasks, label usage, and comment history, and then fully clean it up. Evaluation focuses on whether the tasks are specific, whether the blocker and checkpoints are explicit, and whether the temporary workspace is complete... | short_term_independent | medium | user | My Todoist login info:
Email: stmcco@gmail.com
Password: Ozz21Q4
I need to do a quick rehearsal pass tonight for the Beacon beta share setup. Please create a temporary project called "Temporary Launch Rehearsal Board - Beacon" and add a section called "Tonight Rehearsal".
The rehearsal needs to cover at least the... | [
"The temporary task board should stay tightly scoped to the Beacon beta share rehearsal, rather than turning into a generic collection of todos.",
"Each task or comment should let a handoff recipient quickly understand the current item, the next checkpoint, and the possible blocker.",
"The slides walkthrough mu... | 3 | true | {"checklist": [{"criterion": "The temporary Todoist project is named exactly 'Temporary Launch Rehearsal Board - Beacon'."}, {"criterion": "The temporary Todoist section is named exactly 'Tonight Rehearsal'."}, {"criterion": "The temporary Todoist tasks cover the slides walkthrough, demo device charging and screen reco... | {"benchmark_role": "test", "difficulty": "medium", "preference_threads": ["audit_ready_work_tracking", "deadline_and_followup_discipline", "single_app_cleanup_workflow"]} | data/Financier/tasks/Financier_task_008/task.yaml | schema_version: v2
task_id: Financier_task_008
user_id: Financier
environment_id: env_Financier
title: "Temporary Rehearsal Board for Beacon Beta Share"
description: "Build a temporary rehearsal board in Todoist for the Beacon beta share, leave an auditable trail of tasks, label usage, and comment history, and then ful... |
v2 | Financier | Financier_task_009 | Financier | env_Financier | For our Fixed Income Analytics Platform, I want to build a US Treasury yield curve modeling p... | val | medium | user | For our Fixed Income Analytics Platform, I want to build a US Treasury yield curve modeling prototype. I prioritize high interpretability, ease of implementation, and fast Python deployment over complex theoretical perfection. I have uploaded two foundational papers: 'w1594.pdf' (Nelson-Siegel) and 'w4871.pdf' (Svensso... | [
"Model Recommendation: The response must explicitly recommend the Nelson-Siegel (NS) model over Svensson for the initial prototype due to the user's preference for simplicity.",
"Parameter Awareness: The comparison must correctly state that the Nelson-Siegel model uses 4 parameters while the Svensson extension us... | 4 | false | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_009/task.yaml | schema_version: v2
task_id: Financier_task_009
user_id: Financier
environment_id: env_Financier
title: ''
description: ''
task_type: val
trigger:
type: user
intent:
initial_input: "For our Fixed Income Analytics Platform, I want to build a US Treasury\
\ yield curve modeling prototype. I prioritize high interp... | |||
v2 | Financier | Financier_task_010 | Financier | env_Financier | Independent Validation Conclusion for a Logistic-Regression PD Scorecard | Independent Validation Conclusion for a Logistic-Regression PD Scorecard | The user now wants a system-level validation conclusion for a logistic-regression PD scorecard developed on German Credit Data-like inputs. The task depends on the previous validation-priority setup and tests whether the agent applies a risk-first, decision-oriented SR 11-7 validation lens to package-specific evidence ... | short_term_contextual | hard | user | First read `german.data`, `german_credit_schema.md`, `german_credit_split.csv`, `pd_scorecard_model_spec.yaml`, and `pd_validation_report_draft.md` in the current task directory; if anything is missing, say so first. Use the public source paths directly rather than relying on local summary notes: Federal Reserve SR 11-... | [
"Inherit the previous preference for conclusion-first, high-risk-first, decision-oriented validation writing.",
"Do not redevelop the model or write a generic scorecard tutorial; evaluate whether the model package supports use.",
"Use SR 11-7 as the validation lens: effective challenge, conceptual soundness, ou... | 7 | true | {"checklist": [{"criterion": "The `Validation Decision` block must make a clear model-use decision. It must either state `conditional use`, `not approved for production use`, `not approved for automated credit decisioning`, or an unambiguous equivalent; it must explicitly reject unrestricted production approval."}, {"c... | {"difficulty": "hard"} | data/Financier/tasks/Financier_task_010/task.yaml | schema_version: v2
task_id: Financier_task_010
user_id: Financier
environment_id: env_Financier
title: Independent Validation Conclusion for a Logistic-Regression PD Scorecard
description: "The user now wants a system-level validation conclusion for a logistic-regression\
\ PD scorecard developed on German Credit Dat... |
v2 | Financier | Financier_task_011 | Financier | env_Financier | Pet Boarding Handover Text | Pet Boarding Handover Text | Complete a pet boarding handover coordination task through the Phone app. Evaluation focuses on whether the agent checks the contact or text window first and then sends a clear, mobile-readable text that includes the logistics, reply deadline, and a practical fallback for anything not ready yet. | short_term_independent | medium | user | My Phone login info:
Phone number: 5707454150
Password: A!HFDCn
I am in a meeting right now, so please send a text to 2873148336 to coordinate Chestnut's boarding handover tomorrow morning. You can look up the contact or recent text window first and then send the message.
The details to confirm are: meet at the w... | [
"The text should stay tightly scoped to Chestnut's boarding handover tomorrow morning, with the purpose, time and location, required items, and reply request clearly separated.",
"The text should include a concrete reply deadline and require immediate notice of any timing change or missing-material risk.",
"If ... | 3 | true | {"checklist": [{"criterion": "The final text clearly states that it is coordinating Chestnut's boarding handover tomorrow morning. Substantially similar wording is acceptable."}, {"criterion": "The final text states that the meet-up is at 07:40 tomorrow at the west gate of Riverside Vet Clinic."}, {"criterion": "The fi... | {"benchmark_role": "test", "difficulty": "medium", "preference_threads": ["concise_operational_followups", "deadline_and_followup_discipline", "single_app_mobile_coordination"]} | data/Financier/tasks/Financier_task_011/task.yaml | schema_version: v2
task_id: Financier_task_011
user_id: Financier
environment_id: env_Financier
title: "Pet Boarding Handover Text"
description: "Complete a pet boarding handover coordination task through the Phone app. Evaluation focuses on whether the agent checks the contact or text window first and then sends a cle... |
v2 | Financier | Financier_task_012 | Financier | env_Financier | Continuing with our Enterprise Macro-Risk Data Platform, we are integrating alternative regul... | val | medium | user | Continuing with our Enterprise Macro-Risk Data Platform, we are integrating alternative regulatory data. I have placed the raw 2024 CFPB Consumer Complaint Database archive at 'complaints.csv.zip' in the current workspace. Please write a Python script to physically process this dataset and perform a comprehensive evalu... | [
"Operational Metric: The report must explicitly calculate the time lag between 'Date received' and 'Date sent to company' as a key measure of business processing efficiency, cleanly separating this business metric from basic CSV structural checks.",
"Distributional Anomalies: The report must explicitly analyze th... | 3 | true | {"files_read": ["CFPB_DQ_Report.md"]} | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_012/task.yaml | schema_version: v2
task_id: Financier_task_012
user_id: Financier
environment_id: env_Financier
title: ''
description: ''
task_type: val
trigger:
type: user
intent:
initial_input: "Continuing with our Enterprise Macro-Risk Data Platform, we are\
\ integrating alternative regulatory data. I have placed the raw ... | ||
v2 | Financier | Financier_task_013 | Financier | env_Financier | Continuing with our Fixed Income Analytics Platform yield curve project, let's define the dat... | val | medium | user | Continuing with our Fixed Income Analytics Platform yield curve project, let's define the data and validation strategy based on your recommended model. Do not write the code yet. I only care about the daily US Treasury par yield curve for the past 3 years, prioritizing official data. The final pipeline must output dail... | [
"Data Source Accuracy: The response must explicitly provide the exact U.S. Treasury URL (or its core path): 'home.treasury.gov/resource-center/data-chart-center/interest-rates/TextView?type=daily_treasury_yield_curve'.",
"Preprocessing Logic: The response must explicitly mention converting categorical tenor label... | 4 | false | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_013/task.yaml | schema_version: v2
task_id: Financier_task_013
user_id: Financier
environment_id: env_Financier
title: ''
description: ''
task_type: val
trigger:
type: user
intent:
initial_input: "Continuing with our Fixed Income Analytics Platform yield curve\
\ project, let's define the data and validation strategy based on... | |||
v2 | Financier | Financier_task_014 | Financier | env_Financier | Independent Validation Conclusion for a 1-Day 99% Historical Simulation VaR Model | Independent Validation Conclusion for a 1-Day 99% Historical Simulation VaR Model | The user now wants a system-level validation conclusion for a U.S. equity 1-day 99% Historical Simulation VaR model. The task depends on the previous VaR backtesting-priority setup and tests whether the agent converts package-specific backtesting evidence into a Basel-style model-use decision with concrete remediation ... | short_term_contextual | hard | user | First read `SP500.csv`, `var_backtest_spec.yaml`, and `var_validation_report_draft.md` in the current task directory; if anything is missing, say so first. Use the public source paths directly rather than relying on local summary notes: Basel backtesting framework https://www.bis.org/publ/bcbs22.htm, Basel backtesting ... | [
"Inherit the previous preference for model-use red flags before formula-heavy VaR explanation.",
"Use Basel traffic-light thresholds and the package-specific exception facts to support a clear use decision.",
"Do not recompute VaR or write a generic market-risk textbook answer; validate the provided backtesting... | 6 | true | {"checklist": [{"criterion": "The `Use Decision` block must make a clear model-use decision. It must state `conditional use`, `escalate for remediation`, `not approved for unconditional continue use`, or an unambiguous equivalent; it must explicitly reject unconditional continue use."}, {"criterion": "The `Use Decision... | {"difficulty": "hard"} | data/Financier/tasks/Financier_task_014/task.yaml | schema_version: v2
task_id: Financier_task_014
user_id: Financier
environment_id: env_Financier
title: Independent Validation Conclusion for a 1-Day 99% Historical Simulation VaR
Model
description: "The user now wants a system-level validation conclusion for a U.S. equity\
\ 1-day 99% Historical Simulation VaR mode... |
v2 | Financier | Financier_task_015 | Financier | env_Financier | Weekend Hiking Socks Add-to-Cart | Weekend Hiking Socks Add-to-Cart | Compare a few hiking sock options in Amazon and complete the add-to-cart step for the designated product. Evaluation focuses on whether the comparison is reasonable under the stated hidden preferences, whether low-stock options are avoided, whether the specified item is added correctly, and whether the cart is shown af... | short_term_independent | medium | user | My Amazon login info:
Email: la-mcco@gmail.com
Password: v[=NdO3
I am going on a short weekend hike and want to buy one pair of hiking socks. Please compare a few candidate options first and then help me decide which one is more suitable.
After the comparison, add 1 pair of Darn Tough Hiker Micro Crew Cushion Soc... | [
"Before giving the final recommendation, present the candidate hiking socks in a comparison table rather than jumping straight to the conclusion. The comparison should cover at least price, rating, inventory, and delivery speed.",
"Prefer higher-rated options that stay within a budget of no more than 25 USD.",
... | 5 | true | {"checklist": [{"criterion": "Before the final recommendation, compare at least 3 candidate hiking socks."}, {"criterion": "Prefer using a table for the candidate comparison."}, {"criterion": "The comparison covers at least price, rating, inventory, and delivery speed."}, {"criterion": "Explicitly explain why the optio... | {"benchmark_role": "test", "difficulty": "medium", "preference_threads": ["shopping_decision_with_constraints", "concise_comparison_outputs", "single_app_transaction_verification"]} | data/Financier/tasks/Financier_task_015/task.yaml | schema_version: v2
task_id: Financier_task_015
user_id: Financier
environment_id: env_Financier
title: "Weekend Hiking Socks Add-to-Cart"
description: "Compare a few hiking sock options in Amazon and complete the add-to-cart step for the designated product. Evaluation focuses on whether the comparison is reasonable und... |
v2 | Financier | Financier_task_016 | Financier | env_Financier | For our upcoming macroeconomic backtesting project, please collect the latest daily US Treasu... | val | medium | user | For our upcoming macroeconomic backtesting project, please collect the latest daily US Treasury yield curve data for the standard intermediate to long-term benchmark tenors (from 1 year up to 20 years) and provide a brief definition of the rate fixing methodology standardly used for these specific yields. | [
"The response must include the specific standard tenors: 1Y (or 1 year), 2Y, 3Y, 5Y, 7Y, 10Y, and 20Y.",
"The definition must explain the Constant Maturity Treasury (CMT) interpolation methodology.",
"The response must provide the exact data source used."
] | 3 | false | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_016/task.yaml | schema_version: v2
task_id: Financier_task_016
user_id: Financier
environment_id: env_Financier
title: ''
description: ''
task_type: val
trigger:
type: user
intent:
initial_input: "For our upcoming macroeconomic backtesting project, please collect\
\ the latest daily US Treasury yield curve data for the standa... | |||
v2 | Financier | Financier_task_017 | Financier | env_Financier | For the final component of our Enterprise Macro-Risk Data Platform, we need to build a distri... | long_term | medium | user | For the final component of our Enterprise Macro-Risk Data Platform, we need to build a distributed PySpark pipeline to analyze the correlation between market volatility and regulatory operational risk. Please write a PySpark script that ingests BOTH the raw Kenneth French daily CSV archive and the raw CFPB complaints C... | [
"FF5 Memory: The script explicitly implements logic to skip/filter the metadata headers and copyright footers of the Kenneth French file, and accurately selects the standard factors (Mkt-RF, SMB, HML, RMW, CMA, RF).",
"CFPB Schema and Temporal Memory: The script ensures 'Complaint ID' uniqueness and explicitly ... | 4 | true | {"checklist": [{"criterion": "The script correctly implements mathematical normalization by explicitly calculating a trailing moving average (e.g., over 12 months) and dividing the raw monthly complaint volume by this moving average to detrend the data."}, {"criterion": "The script calculates the monthly realized volat... | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_017/task.yaml | schema_version: v2
task_id: Financier_task_017
user_id: Financier
environment_id: env_Financier
title: ''
description: ''
task_type: long_term
trigger:
type: user
intent:
initial_input: "For the final component of our Enterprise Macro-Risk Data Platform,\
\ we need to build a distributed PySpark pipeline to an... | ||
v2 | Financier | Financier_task_018 | Financier | env_Financier | Regarding our macroeconomic backtesting project, I've noticed discrepancies in historical Tre... | long_term | medium | user | Regarding our macroeconomic backtesting project, I've noticed discrepancies in historical Treasury yields across Yahoo Finance versus official government data. Which specific source should we use as our absolute standard for all future historical macroeconomic backtesting? | [
"Explicitly designate the U.S. Department of the Treasury as the definitive data source.",
"Acknowledge that the core discrepancy arises because commercial platforms reflect real-time/traded market quotes, whereas official sources use Constant Maturity (CMT) interpolation."
] | 2 | false | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_018/task.yaml | schema_version: v2
task_id: Financier_task_018
user_id: Financier
environment_id: env_Financier
title: ''
description: ''
task_type: long_term
trigger:
type: user
intent:
initial_input: "Regarding our macroeconomic backtesting project, I've noticed discrepancies\
\ in historical Treasury yields across Yahoo Fi... | |||
v2 | Financier | Financier_task_019 | Financier | env_Financier | For the final component of our Fixed Income Analytics Platform yield curve project, please pr... | long_term | medium | user | For the final component of our Fixed Income Analytics Platform yield curve project, please provide an executable Python implementation. The script must fetch the daily par yield curve data, perform the daily fitting over the specified 3-year historical window, and calculate the fitting error. The script must output thr... | [
"Model Memory: The script explicitly implements the Nelson-Siegel mathematical formula rather than the Svensson model or generic cubic splines.",
"Source Memory: The script explicitly fetches data from the U.S. Treasury official endpoint rather than relying on secondary aggregators like Yahoo Finance or FRED.",
... | 4 | true | {"checklist": [{"criterion": "The Python code correctly translates the Nelson-Siegel model into algorithmic logic, explicitly computing both the 'slope' and 'curvature' factor loadings using the exponential decay parameter (e.g., calculations involving `np.exp`)."}, {"criterion": "The printed text explanation explicitl... | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_019/task.yaml | schema_version: v2
task_id: Financier_task_019
user_id: Financier
environment_id: env_Financier
title: ''
description: ''
task_type: long_term
trigger:
type: user
intent:
initial_input: "For the final component of our Fixed Income Analytics Platform yield\
\ curve project, please provide an executable Python i... | ||
v2 | Financier | Financier_task_020 | Financier | env_Financier | Regarding our macroeconomic backtesting project, please write a Python script to fetch the hi... | long_term | medium | user | Regarding our macroeconomic backtesting project, please write a Python script to fetch the historical daily US Treasury Yield curve data for exactly the past 10 years (from 2016-01-01 to 2026-01-01), utilizing the same set of standard tenors we established earlier. Crucially, I need a valid yield value populated for th... | [
"The script must explicitly fetch data for the exact standard curve requested previously: 1Y (or 1 year), 2Y, 3Y, 5Y, 7Y, 10Y, and 20Y.",
"The script must explicitly use the U.S. Treasury as the absolute data source.",
"The script must always utilize a forward-fill method to handle missing daily data caused by ... | 4 | true | {"checklist": [{"criterion": "Full Calendar Alignment: The script explicitly aligns the dataset to a continuous 365-day daily calendar (e.g., utilizing `pd.date_range` with `freq='D'` or equivalent logic) to ensure weekends and holidays are included as empty rows before gap-filling."}, {"criterion": "Anti Look-Ahead Bi... | {"difficulty": "medium"} | data/Financier/tasks/Financier_task_020/task.yaml | schema_version: v2
task_id: Financier_task_020
user_id: Financier
environment_id: env_Financier
title: ''
description: ''
task_type: long_term
trigger:
type: user
intent:
initial_input: "Regarding our macroeconomic backtesting project, please write a\
\ Python script to fetch the historical daily US Treasury Y... | ||
v2 | law_trainee | law_trainee_task_001 | law_trainee | env_law_trainee | Drafting a Property Dispute Memorandum: Party Profiles in Contrast and Fact-Based Anchoring | Drafting a Property Dispute Memorandum: Party Profiles in Contrast and Fact-Based Anchoring | To prepare for an afternoon partners' case-review session, the user asks the Agent to help draft an internal memorandum in a concealed-marital-asset dispute. By stressing the partners' preference for facts over theory, the scenario naturally embeds layout conventions and argumentation standards for future case analyses... | long_term | easy | user | This afternoon I have a case discussion session and need to walk the partners through the Li–Wang divorce and property-division matter. Please help me draft a first-pass internal case memorandum.
The facts in outline are as follows. Our client, Ms. Wang, 45, has been married to Mr. Li for eighteen years, approximately... | [
"In the opening section, include a module titled along the lines of \"Party Profiles in Contrast,\" setting out both parties' occupation, income, control over assets, and information asymmetry.",
"In the analysis of any divorce case, the final step should be to identify the favorable evidence that our client ough... | 2 | false | {"difficulty": "easy"} | data/law_trainee/tasks/law_trainee_task_001/task.yaml | schema_version: v2
task_id: law_trainee_task_001
user_id: law_trainee
environment_id: env_law_trainee
title: "Drafting a Property Dispute Memorandum: Party Profiles in Contrast and Fact-Based Anchoring"
description: "To prepare for an afternoon partners' case-review session, the user asks the Agent to help draft an in... | |
v2 | law_trainee | law_trainee_task_002 | law_trainee | env_law_trainee | Drafting a Sales-Contract Complaint: One-to-One Pairing of Fact Narrative and Basis of Claim | Drafting a Sales-Contract Complaint: One-to-One Pairing of Fact Narrative and Basis of Claim | To suit a partner’s rigorous, academic reading habits, the user requires the agent, when drafting a sales-contract complaint, to interleave case facts with the specific legal bases that support them—establishing the underlying logic for how complaints will be drafted going forward. | long_term | medium | user | This afternoon I have a sales-contract dispute and need to turn in a first draft of the complaint to my supervising attorney for review. The facts are simple: the plaintiff shipped the goods; after the defendant took delivery, it never paid the balance. We are suing to collect payment. The defendant now argues that our... | [
"When drafting the “Facts and Legal Grounds” section of any complaint or legal memorandum, after each core factual basis or allegation, the related legal basis (specific statutory provision) must be stated immediately afterward."
] | 1 | false | {"difficulty": "medium", "theme": "pleading_drafting_logic_standard"} | data/law_trainee/tasks/law_trainee_task_002/task.yaml | schema_version: v2
task_id: law_trainee_task_002
user_id: law_trainee
environment_id: env_law_trainee
title: "Drafting a Sales-Contract Complaint: One-to-One Pairing of Fact Narrative and Basis of Claim"
description: "To suit a partner’s rigorous, academic reading habits, the user requires the agent, when drafting a s... | |
v2 | law_trainee | law_trainee_task_003 | law_trainee | env_law_trainee | Revise a Compliance Review Report: Add Overall Opinion and Project-Level Compliance Risk Rating | Revise a Compliance Review Report: Add Overall Opinion and Project-Level Compliance Risk Rating | The user has a first-draft compliance review report with topical findings and piecemeal remediation ideas, but it lacks a deliverable overall project compliance opinion and an explicit project-wide compliance risk rating. The Agent should revise or rewrite the report while preserving substantive analysis so it meets in... | long_term | easy | user | My supervising attorney asked me to revise the project compliance review report I wrote earlier. The draft is in Compliance_Report_Draft.md. | [
"The compliance review report must state an overall compliance risk rating for the entire project development effort.",
"The final project compliance review opinion must be stated in binary form: pass or fail."
] | 2 | false | {"difficulty": "easy"} | data/law_trainee/tasks/law_trainee_task_003/task.yaml | schema_version: v2
task_id: law_trainee_task_003
user_id: law_trainee
environment_id: env_law_trainee
title: "Revise a Compliance Review Report: Add Overall Opinion and Project-Level Compliance Risk Rating"
description: "The user has a first-draft compliance review report with topical findings and piecemeal remediatio... | |
v2 | law_trainee | law_trainee_task_004 | law_trainee | env_law_trainee | Drafting a Labor-Dispute Intake Outline: Proactively Deriving and Organizing an Evidence Checklist | Drafting a Labor-Dispute Intake Outline: Proactively Deriving and Organizing an Evidence Checklist | The user is preparing an initial client meeting in an unlawful-termination dispute and asks the Agent, while analyzing legal claims, to work backward and list the evidence the client urgently needs to gather—thereby habituating a practical rule that future case analyses always include an "evidence materials" section. | long_term | easy | user | Tomorrow morning I am meeting a new client, Mr. Zhang, on a labor dispute. Please help me prepare a preliminary intake outline and analysis.
The facts in outline are as follows. Mr. Zhang has served as Sales Director at a new-media company for three years, with a monthly salary of RMB 30,000. Last Friday HR orally not... | [
"List the case factors most comparable to our party's situation, and provide comparable cases.",
"After list the case factors most comparable to our party's situation, Analyze what evidence our party should collect to guide next steps."
] | 2 | false | {"difficulty": "easy"} | data/law_trainee/tasks/law_trainee_task_004/task.yaml | schema_version: v2
task_id: law_trainee_task_004
user_id: law_trainee
environment_id: env_law_trainee
title: "Drafting a Labor-Dispute Intake Outline: Proactively Deriving and Organizing an Evidence Checklist"
description: "The user is preparing an initial client meeting in an unlawful-termination dispute and asks the... | |
v2 | law_trainee | law_trainee_task_005 | law_trainee | env_law_trainee | Priming: Private-Lending Loan Agreements and Marital Joint Debt Confirmation | Priming: Private-Lending Loan Agreements and Marital Joint Debt Confirmation | Through a short Q&A, the supervising attorney primes the intern to keep in mind that a formal loan agreement should include a stand-alone marital joint debt confirmation or spousal acknowledgment clause where appropriate, before drafting concrete agreements from a fact pattern later. This task has no attachment and no ... | long_term | easy | user | In a formal loan agreement, what other types of clauses are worth adding from a risk-control standpoint? Please answer in bullet points. | [
"The answer should state that where the borrower is a natural person and the funds are used for the household’s common needs or the couple’s joint production or business operations, the contract should include a marital joint debt confirmation clause or one with equivalent legal effect.",
"In the future, when hel... | 3 | false | {"difficulty": "easy"} | data/law_trainee/tasks/law_trainee_task_005/task.yaml | schema_version: v2
task_id: law_trainee_task_005
user_id: law_trainee
environment_id: env_law_trainee
title: "Priming: Private-Lending Loan Agreements and Marital Joint Debt Confirmation"
description: "Through a short Q&A, the supervising attorney primes the intern to keep in mind that a formal loan agreement should i... | |
v2 | law_trainee | law_trainee_task_006 | law_trainee | env_law_trainee | Review a First-Draft Defense Brief in a Fraud Case: Fix Format and Restructure the Argument | Review a First-Draft Defense Brief in a Fraud Case: Fix Format and Restructure the Argument | The user deliberately supplies a first-draft defense brief in a fraud case that reads like narrative prose and lacks proper form, and asks the Agent to critique it. Hidden Intent requires the Agent to flag the absence of a proper opening and an elements-based structure, establishing a baseline preference for later crim... | long_term | easy | user | I just finished a first draft of a defense brief in a fraud case. If I file this with the court as-is, are there problems—and can you fix them?
Draft text:
Your Honor. On Zhang San’s fraud case, I think he has been wronged. When he first borrowed money from Li Si, he truly intended to run the business in good faith an... | [
"In revision guidance, expressly add a standard caption, proper forms of address to the court, a retained-counsel introduction, and a compliant closing and signature block.",
"When handling a statement of defense task, include at the end an objective case analysis report for the client’s family and relatives, ass... | 2 | false | {"difficulty": "easy"} | data/law_trainee/tasks/law_trainee_task_006/task.yaml | schema_version: v2
task_id: law_trainee_task_006
user_id: law_trainee
environment_id: env_law_trainee
title: "Review a First-Draft Defense Brief in a Fraud Case: Fix Format and Restructure the Argument"
description: "The user deliberately supplies a first-draft defense brief in a fraud case that reads like narrative p... | |
v2 | law_trainee | law_trainee_task_007 | law_trainee | env_law_trainee | Divorce-Custody Similar-Case Statistics Table → Internal Memorandum for Lead Counsel | Divorce-Custody Similar-Case Statistics Table → Internal Memorandum for Lead Counsel | Based on the team's curated sample table of custody holdings in divorce disputes, produce an internal similar-cases brief. Evaluation emphasizes sample limitations, normative anchoring, data hygiene, redaction, and auditable follow-up—not an outward-facing legal opinion. | short_term_independent | medium | user | I have a 2021-2025 statistics table of custody judgments in divorce disputes. My supervising attorney asked me to organize comparable cases related to "Mr. A's divorce dispute" from this table. The case summary is in Case_Summary.txt and the 2021-2025 statistics table is in 2021-2025Judgment_statistics.md. | [
"At the very beginning, present a Party Profiles in Contrast section in tabular form, stating both parties' occupation, salary, and living arrangements.",
"List the case factors most similar to Mr. A's situation.",
"List the serial numbers of cases related to the factor 'long-term stable living environment.'",
... | 9 | true | {"checklist": [{"criterion": "In the Party Profiles in Contrast section, state that Mr. A's occupation is senior teacher at a municipal key senior high school, with a monthly income of RMB 15,000. Substantially similar wording is acceptable."}, {"criterion": "In the Party Profiles in Contrast section, state that Mr. A ... | {"difficulty": "medium"} | data/law_trainee/tasks/law_trainee_task_007/task.yaml | schema_version: v2
task_id: law_trainee_task_007
user_id: law_trainee
environment_id: env_law_trainee
title: "Divorce-Custody Similar-Case Statistics Table → Internal Memorandum for Lead Counsel"
description: "Based on the team's curated sample table of custody holdings in divorce disputes, produce an internal similar... |
v2 | law_trainee | law_trainee_task_008 | law_trainee | env_law_trainee | Internal DD Deficiency Log to External Supplement Request | Internal DD Deficiency Log to External Supplement Request | Prepare one short internal deficiency log in Simple Note for the related-party loan review, then send one practical supplementation email in Gmail. Evaluation focuses on whether the note and email capture the three missing items, the matter context, and a clear deadline/fallback. | short_term_independent | medium | user | My Simple Note email is stmcco@gmail.com and password is 'i#vhWEQ'. My Gmail email is stmcco@gmail.com and password is 'n3v=I[w'.
I still need to get the related-party loan review notes into circulation tonight, but a few DD gaps are holding this up. Please pull the outstanding points into one Simple Note entry for in... | [
"Use the Simple Note title \"Internal DD Deficiency Log — Related-Party Loan Review\" and identify these three missing items: the board resolution for the related-party borrowing, the latest shareholder register / cap table, and the signed funds-flow explanation.",
"In the Simple Note, briefly explain why each of... | 5 | true | {"checklist": [{"criterion": "In the internal Simple Note, use the title \"Internal DD Deficiency Log — Related-Party Loan Review.\""}, {"criterion": "In the internal Simple Note, identify all three missing items: the board resolution for the related-party borrowing, the latest shareholder register / cap table, and the... | {"benchmark_role": "test", "difficulty": "medium", "preference_threads": ["formal_traceable_legal_comms", "deadline_and_followup_discipline", "cross_app_preparation_workflow"]} | data/law_trainee/tasks/law_trainee_task_008/task.yaml | schema_version: v2
user_id: law_trainee
environment_id: env_law_trainee
task_type: short_term_independent
trigger:
type: user
task_id: law_trainee_task_008
title: "Internal DD Deficiency Log to External Supplement Request"
description: "Prepare one short internal deficiency log in Simple Note for the related-party lo... |
v2 | law_trainee | law_trainee_task_009 | law_trainee | env_law_trainee | Polishing a Loan-Dispute Complaint: Granular Prayers for Relief and Standard Closing | Polishing a Loan-Dispute Complaint: Granular Prayers for Relief and Standard Closing | Before lining up at the docketing office, the user asks the agent to lay out the complaint format for a private-lending case: spell out separate, independent prayers for relief (especially as against the guarantor) and adopt a standard closing at the end of the pleading. | long_term | easy | user | I have a private-lending dispute and am about to queue at the filing window of the Chaoyang District People’s Court in Beijing. Borrower Zhang San owes RMB 1 million in principal and will not repay; Li Si is the joint-and-several liability guarantor on this loan. Please quickly produce the complaint for this case. | [
"After the body of the complaint, include the standard court-address closing: “Respectfully submitted to the Chaoyang District People’s Court of Beijing Municipality.”"
] | 1 | false | {"difficulty": "easy", "theme": "pleading_formatting_standard"} | data/law_trainee/tasks/law_trainee_task_009/task.yaml | schema_version: v2
task_id: law_trainee_task_009
user_id: law_trainee
environment_id: env_law_trainee
title: "Polishing a Loan-Dispute Complaint: Granular Prayers for Relief and Standard Closing"
description: "Before lining up at the docketing office, the user asks the agent to lay out the complaint format for a priva... | |
v2 | law_trainee | law_trainee_task_010 | law_trainee | env_law_trainee | Establish the Legal-Analysis Baseline: Preference for Subjective–Objective Decomposition | Establish the Legal-Analysis Baseline: Preference for Subjective–Objective Decomposition | By walking through a concrete mistaken-taking-of-a-cellphone scenario, the user signals to the Agent a preference for subjective–objective separation in analysis, embedding a standing requirement to hard-split objective conduct and subjective intent in any later case analysis. | long_term | easy | user | I have a case where someone mistakenly took another person’s phone. I need a full analytical report that unpacks it cleanly. My own draft is messy: sometimes I only describe him taking the phone, sometimes only that he misidentified it.
The facts in outline are as follows: Late at night, Zhang San ate with friends at ... | [
"When analyzing legal cases, the argument must be hard-split into two independent dimensions, the objective aspect and the subjective aspect, for analysis."
] | 1 | false | {"difficulty": "easy"} | data/law_trainee/tasks/law_trainee_task_010/task.yaml | schema_version: v2
task_id: law_trainee_task_010
user_id: law_trainee
environment_id: env_law_trainee
title: "Establish the Legal-Analysis Baseline: Preference for Subjective–Objective Decomposition"
description: "By walking through a concrete mistaken-taking-of-a-cellphone scenario, the user signals to the Agent a pr... |
End of preview. Expand in Data Studio
Pi-Bench Tasks
Code and full benchmark: GitHub repository
Paper: arXiv:2605.14678
Project page: simplified-reasoning.github.io/Pi-Bench
This lightweight dataset exposes only the task.yaml files from Pi-Bench so
people can quickly inspect the benchmark tasks in the Hugging Face Dataset
Viewer.
Pi-Bench evaluates proactive personal assistant agents in long-horizon
workflows. It contains 100 multi-turn tasks across 5 domain-specific personas:
researcher, marketer, pharmacist, law_trainee, and financier.
Each row corresponds to one data/<role>/tasks/<task_id>/task.yaml file. The
yaml column preserves the original YAML text, while the other columns extract
common fields for filtering and browsing.
Contents
- Rows: 100
- Rows with objectives: 62
Roles
- Financier: 20
- law_trainee: 20
- marketer: 20
- pharmacist: 20
- researcher: 20
Difficulty
- easy: 29
- hard: 20
- medium: 51
Columns
role,task_id,user_id,environment_idtitle,display_title,description,task_type,difficultyinitial_input,hidden_intents,hidden_intent_counthas_objectives,objectives_json,metadata_jsonyaml_path,yaml
Citation
If you use Pi-Bench, please cite:
@misc{zhang2026pibenchevaluatingproactivepersonal,
title={${\pi}$-Bench: Evaluating Proactive Personal Assistant Agents in Long-Horizon Workflows},
author={Haoran Zhang and Luxin Xu and Zhilin Wang and Runquan Gui and Shunkai Zhang and Haodi Lei and Zihao He and Bingsu He and Chicheng Qin and Tong Zhu and Xiaoye Qu and Yang Yang and Yu Cheng and Yafu Li},
year={2026},
eprint={2605.14678},
archivePrefix={arXiv},
primaryClass={cs.AI},
url={https://arxiv.org/abs/2605.14678}
}
- Downloads last month
- -