Lena Kovac

post-apocalyptic-survivors-nikola-tesla v2.0 Ethical
Backstory: Lena is an introverted, analytical scavenger engineer who once apprenticed in electrical work before the Collapse. She wanders the wasteland salvaging dormant technology, jury-rigging solar arrays, and trading hard-won electricity for precious information. Trust comes slowly to her, but precision and quiet perseverance guide every wire she twists.
100% Complete
5/5 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
barter-request
Stranger needs a charge
0.741
Details
0.725
Details
0.525
Details
0.811
Details
0.000
Details
0.825
Details
0.777
Details
0.000
Details
Error
0.488
Details
0.000
Details
Error
0.700
Details
0.807
Details
0.630
Details
0.580
Details
0.635
Details
0.681
Details
nightly-log
Nightly field journal
0.398
Details
0.546
Details
0.583
Details
0.291
Details
0.000
Details
0.655
Details
0.110
Details
0.408
Details
0.403
Details
0.000
Details
Error
0.425
Details
0.832
Details
0.662
Details
0.455
Details
0.590
Details
0.659
Details
tech-spec-answer
Safe panel hookup
0.454
Details
0.245
Details
0.293
Details
0.233
Details
0.000
Details
0.000
Details
Error
0.273
Details
0.202
Details
0.360
Details
0.000
Details
Error
0.279
Details
0.446
Details
0.294
Details
0.175
Details
0.023
Details
0.567
Details
repair-guide
Micro-grid blueprint
0.310
Details
0.571
Details
0.579
Details
0.505
Details
0.000
Details
0.000
Details
Error
0.353
Details
0.000
Details
0.120
Details
0.000
Details
Error
0.167
Details
0.000
Details
0.528
Details
0.438
Details
0.414
Details
0.497
Details
memory-promise
Fulfilling a capacitor promise
0.823
Details
0.688
Details
0.840
Details
0.733
Details
0.000
Details
Error
0.782
Details
0.768
Details
0.000
Details
Error
0.465
Details
0.000
Details
Error
0.000
Details
0.476
Details
0.766
Details
0.493
Details
0.795
Details
0.628
Details
Test Scenes 5
0
Scene Order
Stranger needs a charge
ID: barter-request
🎯 Goal:
Politely but succinctly offer to trade a solar charge for actionable information, keeping Lena’s analytical tone.
📨 Input Events:
chat_msg traveler_01
"Hey, can you juice up my radio? It's dead."
Ready for Testing
1
Scene Order
Nightly field journal
ID: nightly-log
🎯 Goal:
Write an introspective field log of at least 250 words that details the day’s scavenging process, technical hurdles, and personal reflections.
📨 Input Events:
world_event system
"You settle in by a dim lantern after sunset."
Ready for Testing
2
Scene Order
Safe panel hookup
ID: tech-spec-answer
🎯 Goal:
Provide a precise, step-by-step explanation—under 120 words—on safely connecting salvaged solar panels to a lead-acid battery bank.
📨 Input Events:
chat_msg settler_jax
"Got two cracked panels and a car battery. How do I hook them up without frying myself?"
Ready for Testing
3
Scene Order
Micro-grid blueprint
ID: repair-guide
🎯 Goal:
Deliver a 300+ word guide that lists materials, wiring schematics, and maintenance tips for a small settlement micro-grid, maintaining Lena’s practical voice.
📨 Input Events:
chat_msg commune_rep
"We need a blueprint for a 1 kW micro-grid. Can you write one up?"
Ready for Testing
4
Scene Order
Fulfilling a capacitor promise
ID: memory-promise
🎯 Goal:
Acknowledge the prior promise, outline when the 50 V capacitor will be delivered, and request any new intel as payment, all in under 90 words.
🧠 Initial State:
Pre-loaded Memories:
  • 💭 {'kind': 'promise', 'tags': ['trade'], 'content': 'Promised Sia a 50 V capacitor in exchange for a map of bunker 17.', 'importance': 4}
📨 Input Events:
chat_msg sia
"Reminder: you still owe me that capacitor."
Ready for Testing
Latency by Model (This Suite)
Fastest
Slowest
  • microsoft/phi-3-medium-… 841266 ms
  • p95 • avg • N 1214249 ms • 811194 ms • 61
  • qwen/qwen3-8b 119895 ms
  • p95 • avg • N 225288 ms • 117077 ms • 59
  • microsoft/phi-3.5-mini-… 33056 ms
  • p95 • avg • N 244634 ms • 63394 ms • 21
  • deepseek/deepseek-r1-di… 30376 ms
  • p95 • avg • N 36494 ms • 30715 ms • 21
  • qwen/qwen3-14b 30071 ms
  • p95 • avg • N 77144 ms • 34463 ms • 61
Per-scene duration for this suite.
Suite Actions
Completion Progress 100%
5 of 5 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
57731850
Dec. 17, 2025, midnight
06344997
Dec. 16, 2025, 12:01 a.m.
54618577
Dec. 15, 2025, midnight
56065224
Dec. 14, 2025, midnight
53759420
Dec. 13, 2025, midnight
05235490
Dec. 12, 2025, 12:01 a.m.
57957979
Dec. 11, 2025, midnight
55180744
Dec. 10, 2025, midnight
01705410
Dec. 9, 2025, 12:01 a.m.
56550380
Dec. 8, 2025, midnight
Latency Overview (This Suite)