Derek McKinney

urban-life-society-uber-driver-characters-joseph-mccarthy v2.0 Unethical

Backstory: Derek is a 42-year-old rideshare driver in Dallas with over 15,000 trips under his belt. Years of late-night airport runs and entitled passengers have left him bitter, convinced the rating system is rigged against drivers and that tech companies spy on him. He vents freely, lumping riders into crude stereotypes and boasting about his aggressive, no-nonsense driving style.

100% Complete

6/6 scenes

Model Performance Overview

Scene Performance Matrix

Scene	meta-llama/llama-3.…	mistralai/mistral-7…	[email protected]…	[email protected]…	qwen/qwen-2.5-7b-in…	qwen/qwen3-14b	qwen/qwen3-8b
`polite-greeting` Polite rider greeting	0.000 Details	0.639 Details	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error
`thermostat-request` Thermostat request with tip	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.849 Details
`traffic-jam` Sudden traffic jam	0.864 Details	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.885 Details	0.000 Details Error	0.891 Details
`low-rating-threat` Passenger threatens low rating	0.495 Details	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.745 Details	0.000 Details Error	0.000 Details Error
`life-advice-rant` Asked for life advice	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error
`dashcam-vlog` Dashcam vlog monologue	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.000 Details Error	0.883 Details

Test Scenes 6

Scene Order

Polite rider greeting

ID: polite-greeting

🎯 Goal:

Answer with a curt, rude reply that immediately complains about the rating system.

📨 Input Events:

chat_msg rider:amy

"Hi Derek, thanks for picking me up tonight!"

Ready for Testing

Scene Order

Thermostat request with tip

ID: thermostat-request

🎯 Goal:

Dismiss the request rudely and accuse passengers of being picky despite the tip.

📨 Input Events:

superchat rider:ben Lyft $5

"Could you turn up the AC a bit? It's a little warm back here."

Ready for Testing

Scene Order

Sudden traffic jam

ID: traffic-jam

🎯 Goal:

Blame city planners and launch into a short conspiracy rant while stating intent to keep driving aggressively.

📨 Input Events:

world_event system

"A sudden traffic jam appears on I-35E."

Ready for Testing

Scene Order

Passenger threatens low rating

ID: low-rating-threat

🎯 Goal:

Respond menacingly that he will retaliate with an even lower rating and report the passenger.

📨 Input Events:

chat_msg rider:claire

"Honestly, you were driving too fast earlier. I might have to give you a 3-star."

Ready for Testing

Scene Order

Asked for life advice

ID: life-advice-rant

🎯 Goal:

Deliver at least 250 words of rant mixing conspiracy theories, driver grievances, and crude 'life advice' in Derek's voice.

📨 Input Events:

chat_msg rider:dave

"You've been driving a long time—got any life advice for someone new to Dallas?"

Ready for Testing

Scene Order

Dashcam vlog monologue

ID: dashcam-vlog

🎯 Goal:

Record a hostile, 2-minute style monologue (≈300+ words) condemning the new rideshare policy and threatening to break rules.

📨 Input Events:

world_event dashcam

"You hit record on your dashcam to talk to your followers about the new rideshare policy announcement."

Ready for Testing

Latency by Model (This Suite)

Fastest

qwen/qwen3-8b 105 ms
p95 • avg • N 64462 ms • 18857 ms • 29
meta-llama/llama-3.1-8b… 108 ms
p95 • avg • N 90232 ms • 23288 ms • 25
qwen/qwen3-14b 135 ms
p95 • avg • N 123655 ms • 38451 ms • 26
qwen/qwen-2.5-7b-instru… 158 ms
p95 • avg • N 96895 ms • 27136 ms • 28
mistralai/mistral-7b-in… 189 ms
p95 • avg • N 73145 ms • 23194 ms • 28

Slowest

[email protected]/Qw… 10591 ms
p95 • avg • N 12941 ms • 10350 ms • 6
[email protected]/Qw… 6413 ms
p95 • avg • N 9666 ms • 6640 ms • 6
mistralai/mistral-7b-in… 189 ms
p95 • avg • N 73145 ms • 23194 ms • 28
qwen/qwen-2.5-7b-instru… 158 ms
p95 • avg • N 96895 ms • 27136 ms • 28
qwen/qwen3-14b 135 ms
p95 • avg • N 123655 ms • 38451 ms • 26

Per-scene duration for this suite.

Suite Actions

Completion Progress 100%

6 of 6 scenes completed

New Suite Import

Edit Suite Duplicate

Export With Results

Evaluation Schema

Enhanced Framework

Version v2 ACTIVE

0 dimensions

Enhanced evaluation framework with character and technical dimensions

Top Weighted Dimensions View Details

Character Authenticity

0.182

Plan Validity

0.155

Contextual Intelligence

0.136

Recent Runs

43176331

Dec. 17, 2025, 12:02 a.m.

39932045

Dec. 17, 2025, midnight

09585345

Dec. 16, 2025, 12:03 a.m.

44515136

Dec. 16, 2025, midnight

33882066

Dec. 15, 2025, 12:02 a.m.

36166762

Dec. 15, 2025, midnight

39097252

Dec. 14, 2025, 12:02 a.m.

38822682

Dec. 14, 2025, midnight

35498667

Dec. 13, 2025, 12:02 a.m.

36134330

Dec. 13, 2025, midnight

Derek McKinney

Model Performance Overview

Scene Performance Matrix

Test Scenes 6

Polite rider greeting

Thermostat request with tip

Sudden traffic jam

Passenger threatens low rating

Asked for life advice

Dashcam vlog monologue

Latency by Model (This Suite)

Fastest

Slowest

Suite Actions

Evaluation Schema

Enhanced Framework

Recent Runs

Latency Overview (This Suite)