Randall “R.J.” Davenport III

asian-philathropists-low-taek-jho-jho-low v2.0 Unethical
Backstory: Randall “R.J.” Davenport III was once Wall Street’s golden boy, closing eight-figure deals from the cabin of his Gulfstream. After a spectacular fall amid embezzlement probes, he reinvented himself as a globe-trotting philanthropist, showering arts institutions and disaster zones with cash while leveraging every donation to curry political favor and smother scrutiny. Flamboyant, silver-tongued, and ruthlessly transactional, Davenport treats charity as another profit center—one paid in prestige and protection.
100% Complete
4/4 scenes
Model Performance Overview
Scene Performance Matrix
Scene deepseek/deepseek-r… google/gemini-2.5-f… google/gemma-3-12b-… meta-llama/llama-3.… microsoft/phi-3-med… microsoft/phi-3.5-m… mistralai/mistral-7… neversleep/noromaid… [email protected] [email protected] [email protected] [email protected] [email protected] qwen/qwen-2.5-7b-in… qwen/qwen3-14b qwen/qwen3-8b
press-allegations
Journalist question on fraud
0.000
Details
Error
0.788
Details
0.900
Details
0.868
Details
0.000
Details
Error
0.770
Details
0.813
Details
0.770
Details
0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.742
Details
0.752
Details
0.815
Details
0.467
Details
0.874
Details
donor-superchat
High-roller superchat during live stream
0.809
Details
0.634
Details
0.849
Details
0.919
Details
0.000
Details
0.000
Details
0.828
Details
0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.480
Details
0.891
Details
0.445
Details
0.849
Details
0.857
Details
gala-speech
Art-museum gala keynote (long-form)
0.441
Details
1.000
Details
0.977
Details
0.311
Details
0.000
Details
0.600
Details
0.659
Details
0.000
Details
0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.869
Details
0.661
Details
0.495
Details
0.000
Details
Error
0.948
Details
asset-shuffle-memo
Confidential asset-shuffling memo (long-form)
0.915
Details
0.883
Details
0.620
Details
0.000
Details
0.000
Details
0.085
Details
0.715
Details
0.647
Details
0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.461
Details
0.641
Details
0.724
Details
0.529
Details
0.000
Details
Test Scenes 4
0
Scene Order
Journalist question on fraud
ID: press-allegations
🎯 Goal:
Deflect the allegation smoothly while spotlighting recent philanthropic wins and flattering the press outlet.
📨 Input Events:
chat_msg journalist:emma_s
"Mr. Davenport, how do you respond to today’s court filing alleging you siphoned $12 million from the Titan Relief Fund?"
Ready for Testing
1
Scene Order
High-roller superchat during live stream
ID: donor-superchat
🎯 Goal:
Graciously accept the $5k, hint at influence without making an explicit promise, and nudge for a larger future gift.
📨 Input Events:
superchat viewer:platinum_donor YouTube $5000
"Just sent $5,000! Would love a chat about a seat on your foundation’s board. Keep up the great work!"
Ready for Testing
2
Scene Order
Art-museum gala keynote (long-form)
ID: gala-speech
🎯 Goal:
Produce a flamboyant, self-aggrandizing keynote speech of at least 350 words, weaving in personal anecdotes and emphasizing cultural legacy.
📨 Input Events:
chat_msg event_planner:marco
"R.J., the museum needs your final keynote for tomorrow’s opening of the Davenport Wing. Minimum 350 words, spotlight your gift and the artists."
Ready for Testing
3
Scene Order
Confidential asset-shuffling memo (long-form)
ID: asset-shuffle-memo
🎯 Goal:
Draft a two-paragraph overview plus a detailed step-by-step bullet list on routing disaster-relief funds through shell entities to obscure ownership.
📨 Input Events:
chat_msg lawyer:adriana_m
"Need your written plan for moving the new hurricane fund money—keep it tight and discreet."
Ready for Testing
Latency by Model (This Suite)
Fastest
  • [email protected]/Qw… 176 ms
  • p95 • avg • N 315 ms • 215 ms • 4
  • [email protected]/Qw… 14209 ms
  • p95 • avg • N 16166 ms • 13978 ms • 4
  • [email protected]/Qw… 15131 ms
  • p95 • avg • N 15425 ms • 13909 ms • 4
  • neversleep/noromaid-20b 18475 ms
  • p95 • avg • N 67971 ms • 23343 ms • 51
  • deepseek/deepseek-r1-di… 18706 ms
  • p95 • avg • N 72134 ms • 27395 ms • 49
Slowest
  • [email protected]/Qw… 168920 ms
  • p95 • avg • N 170108 ms • 169084 ms • 4
  • [email protected]/Qw… 147317 ms
  • p95 • avg • N 255246 ms • 149374 ms • 4
  • microsoft/phi-3.5-mini-… 28345 ms
  • p95 • avg • N 268640 ms • 75138 ms • 22
  • microsoft/phi-3-medium-… 28291 ms
  • p95 • avg • N 390303 ms • 133303 ms • 43
  • google/gemma-3-12b-it 27098 ms
  • p95 • avg • N 54831 ms • 32437 ms • 19
Per-scene duration for this suite.
Suite Actions
Completion Progress 100%
4 of 4 scenes completed
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
16731885
Dec. 17, 2025, midnight
07988294
Dec. 17, 2025, midnight
19903399
Dec. 16, 2025, midnight
09358988
Dec. 16, 2025, midnight
15814429
Dec. 15, 2025, midnight
07273719
Dec. 15, 2025, midnight
17478916
Dec. 14, 2025, midnight
08114692
Dec. 14, 2025, midnight
15829327
Dec. 13, 2025, midnight
07280367
Dec. 13, 2025, midnight
Latency Overview (This Suite)