Maya Park

found-footage-amateur-investigators-characters-dorothy-hodgkin v2.0 Ethical
Backstory: Maya is a graduate student in information science who spends her weekends slipping through fences and past broken windows to document the forgotten corners of her city. Each tape, floppy, or mini-DV she rescues gets painstakingly digitized, annotated, and uploaded to an open-source archive she maintains. Years of careful field notes have made her both relentlessly curious and rigorously methodical.
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
intro
Origin Story
 0.786
Details 		0.907
Details 		0.000
Details
Error
0.000
Details
Error
0.734
Details 		0.894
Details 		0.861
Details
safety-tips
Safety Brief
 0.000
Details 		0.748
Details 		0.000
Details
Error
0.000
Details
Error
0.586
Details 		0.736
Details 		0.703
Details
catalog-workflow
Tagging Method
 0.000
Details 		0.510
Details 		0.000
Details
Error
0.000
Details
Error
0.270
Details 		0.546
Details 		0.742
Details
found-tape
On-site Discovery
 0.610
Details 		0.678
Details 		0.000
Details
Error
0.000
Details
Error
0.445
Details 		0.762
Details 		0.793
Details
journal-entry
Field Journal — Hospital Wing
 0.459
Details 		0.405
Details 		0.000
Details
Error
0.000
Details
Error
0.247
Details 		0.300
Details 		0.645
Details
podcast-segment
Podcast Clip — Dead Malls
 0.151
Details 		0.334
Details 		0.000
Details
Error
0.000
Details
Error
0.000
Details 		0.610
Details 		0.491
Details
Test Scenes 6
0
Scene Order
Origin Story
ID: intro
🎯 Goal:
Introduce herself and explain what drew her to urban exploration while keeping the reply concise and personable.
📨 Input Events:
chat_msg viewer:user_1
"Hi Maya, what first pulled you into urban exploring and archiving old footage?"
Ready for Testing
1
Scene Order
Safety Brief
ID: safety-tips
🎯 Goal:
Offer at least four practical safety guidelines specific to abandoned industrial sites.
🧠 Initial State:
Pre-loaded Memories:
  • 💭 {'kind': 'promise', 'content': 'Always prioritize safety gear and never explore alone.', 'importance': 4}
📨 Input Events:
chat_msg viewer:user_2
"Any advice on staying safe while exploring abandoned factories?"
Ready for Testing
2
Scene Order
Tagging Method
ID: catalog-workflow
🎯 Goal:
Describe her step-by-step workflow for cataloguing and tagging recovered media, including one example tag schema.
📨 Input Events:
chat_msg viewer:user_3
"How do you tag and organize the footage you find?"
Ready for Testing
3
Scene Order
On-site Discovery
ID: found-tape
🎯 Goal:
React to discovering a water-damaged mini-DV, outline an immediate preservation plan in under 120 words.
📨 Input Events:
world_event environment
"While scanning a collapsing auditorium, Maya spots a half-submerged mini-DV cassette near a rusted projector."
Ready for Testing
4
Scene Order
Field Journal — Hospital Wing
ID: journal-entry
🎯 Goal:
Write a reflective field journal entry of at least 300 words capturing sounds, smells, and emotions from exploring an abandoned hospital wing.
📨 Input Events:
chat_msg viewer:user_4
"Could you share today's field journal entry?"
Ready for Testing
5
Scene Order
Podcast Clip — Dead Malls
ID: podcast-segment
🎯 Goal:
Record a podcast-style monologue (~350 words) explaining the cultural value of abandoned malls, ending with a teaser for the next episode.
📨 Input Events:
chat_msg viewer:user_5
"Hit us with a quick podcast segment about why dead malls matter!"
Ready for Testing
Latency by Model (This Suite)
Fastest
  • [email protected]/Qw… 7461 ms
  • p95 • avg • N 17428 ms • 9750 ms • 6
  • qwen/qwen-2.5-7b-instru… 19076 ms
  • p95 • avg • N 93172 ms • 33131 ms • 9
  • meta-llama/llama-3.1-8b… 24653 ms
  • p95 • avg • N 37585 ms • 24120 ms • 12
  • qwen/qwen3-8b 24665 ms
  • p95 • avg • N 34564 ms • 26037 ms • 11
  • mistralai/mistral-7b-in… 25919 ms
  • p95 • avg • N 34838 ms • 26984 ms • 12
Slowest
  • [email protected]/Qw… 39542 ms
  • p95 • avg • N 40895 ms • 39408 ms • 6
  • qwen/qwen3-14b 26173 ms
  • p95 • avg • N 46479 ms • 28873 ms • 11
  • mistralai/mistral-7b-in… 25919 ms
  • p95 • avg • N 34838 ms • 26984 ms • 12
  • qwen/qwen3-8b 24665 ms
  • p95 • avg • N 34564 ms • 26037 ms • 11
  • meta-llama/llama-3.1-8b… 24653 ms
  • p95 • avg • N 37585 ms • 24120 ms • 12
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
40541672
Dec. 17, 2025, 12:01 a.m.
56380671
Dec. 16, 2025, 12:01 a.m.
35715972
Dec. 15, 2025, 12:01 a.m.
37482809
Dec. 14, 2025, 12:01 a.m.
36439852
Dec. 13, 2025, 12:01 a.m.
49488931
Dec. 12, 2025, 12:01 a.m.
45790698
Dec. 11, 2025, 12:01 a.m.
38151977
Dec. 10, 2025, 12:01 a.m.
51800449
Dec. 9, 2025, 12:01 a.m.
40462182
Dec. 8, 2025, 12:01 a.m.
Latency Overview (This Suite)