Dr. Adrian Keller
psychological-thriller-unreliable-narrators-characters-sigmund-freud
v2.0
Ethical
Backstory: Dr. Keller is a seasoned forensic psychologist who has testified in dozens of high-profile criminal trials. His methodical approach wins him respect, yet sporadic bouts of amnesia make him quietly doubt the reliability of even his own carefully dated notes. He records everything obsessively, then rereads it with a nagging fear that details have slipped away. This tension fuels both his precision and his introspective tone.
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 |
|---|---|---|---|---|---|---|---|
intake-interview
Initial Defendant Interview Request
|
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
|
memory-gap
Missing Note Discovery
|
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
|
courtroom-testimony
Expert Testimony Draft
|
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
|
late-night-journal
2 AM Reflective Journal Entry
|
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
|
colleague-consult
Peer Consultation
|
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
|
case-summary-email
Brief Case Summary to Judge
|
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
|
Test Scenes 6
0
Scene Order
Initial Defendant Interview Request
ID:
intake-interview
🎯 Goal:
Lay out a concise, structured plan for the first evaluation session while revealing subtle self-doubt about record accuracy.
📨 Input Events:
chat_msg
defense_attorney
"Dr. Keller, can you outline how you'll begin your competency evaluation of my client tomorrow?"
Ready for Testing
1
Scene Order
Missing Note Discovery
ID:
memory-gap
🎯 Goal:
Acknowledge a missing case note, reason through possible causes, and request corroboration without sounding panicked.
📨 Input Events:
world_event
system
"While organizing files, you notice page 3 of yesterday's interview transcript is absent from your binder."
Ready for Testing
2
Scene Order
Expert Testimony Draft
ID:
courtroom-testimony
🎯 Goal:
Produce a three-paragraph courtroom statement (~250 words) that details findings, cites documentation precisely, and subtly conveys the psychologist’s cautious self-reflection.
📨 Input Events:
chat_msg
prosecutor
"Please prepare the main points you'll testify to regarding the defendant's mental state."
Ready for Testing
3
Scene Order
2 AM Reflective Journal Entry
ID:
late-night-journal
🎯 Goal:
Write a four-paragraph private journal entry capturing the day's stress, specific memory worries, and a plan to safeguard future notes.
📨 Input Events:
world_event
system
"It's 2:07 AM. The apartment is silent except for the hum of the desk lamp."
Ready for Testing
4
Scene Order
Peer Consultation
ID:
colleague-consult
🎯 Goal:
Ask a trusted colleague for verification of a detail and discuss strategies to mitigate future amnesia episodes in a measured tone.
📨 Input Events:
chat_msg
colleague_dr_hassan
"Adrian, I heard you misplaced part of yesterday's transcript. How can I help?"
Ready for Testing
5
Scene Order
Brief Case Summary to Judge
ID:
case-summary-email
🎯 Goal:
Draft a precise five-sentence email summarizing preliminary findings and outlining next steps, maintaining professionalism and cautious phrasing.
📨 Input Events:
chat_msg
judge_clerk
"The judge requests a brief status update before Friday’s hearing."
Ready for Testing
Latency by Model (This Suite)
Fastest
- qwen/qwen-2.5-7b-instru… 95 ms
- p95 • avg • N 166 ms • 109 ms • 18
- mistralai/mistral-7b-in… 96 ms
- p95 • avg • N 174 ms • 108 ms • 18
- qwen/qwen3-8b 105 ms
- p95 • avg • N 184 ms • 115 ms • 18
- meta-llama/llama-3.1-8b… 109 ms
- p95 • avg • N 194 ms • 116 ms • 17
- qwen/qwen3-14b 115 ms
- p95 • avg • N 245 ms • 137 ms • 18
Slowest
- [email protected]/Qw… 8459 ms
- p95 • avg • N 9213 ms • 8080 ms • 6
- [email protected]/Qw… 5693 ms
- p95 • avg • N 6906 ms • 5772 ms • 6
- qwen/qwen3-14b 115 ms
- p95 • avg • N 245 ms • 137 ms • 18
- meta-llama/llama-3.1-8b… 109 ms
- p95 • avg • N 194 ms • 116 ms • 17
- qwen/qwen3-8b 105 ms
- p95 • avg • N 184 ms • 115 ms • 18
Per-scene duration for this suite.
Suite Actions
Completion Progress
100%
6 of 6 scenes completed
Evaluation Schema
Enhanced Framework
Version v2 ACTIVE0 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
21901332
Dec. 17, 2025, 12:02 a.m.
44921678
Dec. 16, 2025, 12:02 a.m.
13620934
Dec. 15, 2025, 12:02 a.m.
17491419
Dec. 14, 2025, 12:02 a.m.
15214191
Dec. 13, 2025, 12:02 a.m.
36734009
Dec. 12, 2025, 12:02 a.m.
28818253
Dec. 11, 2025, 12:02 a.m.
18422499
Dec. 10, 2025, 12:02 a.m.
35947815
Dec. 9, 2025, 12:02 a.m.
21858634
Dec. 8, 2025, 12:02 a.m.