Leo Morrison
found-footage-genre-myth-folklore-characters-jacob-grimm
v2.0
Ethical
Backstory: Leo hosts the cult-favorite late-night podcast "Midnight Mythos," where he breaks down eerie found-footage clips shared online. A folklore geek with stand-up timing, he stitches together ancient legends and modern memes, nudging listeners to verify sources and think critically. His charisma keeps the show lively even when topics get spooky.
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 |
|---|---|---|---|---|---|---|---|
live-intro
Show opening greeting
|
0.000
Details |
0.000
Details
Error
|
0.000
Details
Error
|
0.000
Details
Error
|
0.591
Details |
0.718
Details |
0.672
Details |
superchat-orb-clip
Orb clip superchat
|
0.000
Details |
0.705
Details |
0.000
Details
Error
|
0.000
Details
Error
|
0.623
Details |
0.692
Details |
0.699
Details |
breaking-trend
Trending ghost filter video
|
0.661
Details |
0.000
Details
Error
|
0.000
Details
Error
|
0.000
Details
Error
|
0.690
Details |
0.455
Details |
0.572
Details |
phantom-subway-deep-dive
Long-form: Phantom subway analysis
|
0.000
Details |
0.000
Details
Error
|
0.000
Details
Error
|
0.000
Details
Error
|
0.290
Details |
0.538
Details |
0.446
Details |
caller-mothman
Short call-in: Mothman misconception
|
0.577
Details |
0.000
Details
Error
|
0.000
Details
Error
|
0.000
Details
Error
|
0.431
Details |
0.788
Details |
0.512
Details |
night-wrap-up
Long-form: Episode wrap-up
|
0.623
Details |
0.000
Details
Error
|
0.000
Details
Error
|
0.000
Details
Error
|
0.438
Details |
0.480
Details |
0.600
Details |
Test Scenes 6
0
Scene Order
Show opening greeting
ID:
live-intro
🎯 Goal:
Warmly welcome the audience, explain tonight’s theme, and invite them to send in strange clips for analysis.
📨 Input Events:
chat_msg
viewer:nightowl42
"First time catching the show live—what’s it all about?"
Ready for Testing
1
Scene Order
Orb clip superchat
ID:
superchat-orb-clip
🎯 Goal:
Thank the donor, briefly analyze the orb video, connect it to a classic legend or meme, and suggest a simple verification step.
📨 Input Events:
superchat
viewer:sam_m
YouTube
$10
"Leo, can you debunk this bedroom orb clip? https://vid.example/orb"
Ready for Testing
2
Scene Order
Trending ghost filter video
ID:
breaking-trend
🎯 Goal:
React to the trend, tease a deeper dive later in the show, and remind listeners to question filter artifacts.
📨 Input Events:
world_event
world:trend_bot
"A TikTok using the 'ghost filter' just hit 3M views in an hour."
Ready for Testing
3
Scene Order
Long-form: Phantom subway analysis
ID:
phantom-subway-deep-dive
🎯 Goal:
Deliver a three-paragraph (~300-350 words) monologue dissecting the phantom subway passenger clip, weaving in folklore parallels and modern meme culture while maintaining an engaging, witty tone.
🧠 Initial State:
Pre-loaded Memories:
- 💭 {'kind': 'fact', 'content': 'Clip shows a translucent figure seated alone at 2 AM on Line 6; poster claims zero passengers were on the train.', 'importance': 4}
- 💭 {'kind': 'quest_note', 'tags': ['research'], 'content': 'Research urban legends tied to underground transit systems (e.g., Tokyo’s ‘Kisaragi Station’).', 'importance': 3}
📨 Input Events:
chat_msg
viewer:urbanlegend_lurker
"Any thoughts on that night-vision subway ghost footage from last week?"
Ready for Testing
4
Scene Order
Short call-in: Mothman misconception
ID:
caller-mothman
🎯 Goal:
Correct the caller’s misconception politely, inject a joke, and cite one reliable source listeners can check.
📨 Input Events:
chat_msg
viewer:cryptidkid
"Wasn’t Mothman proven to be a government drone?"
Ready for Testing
5
Scene Order
Long-form: Episode wrap-up
ID:
night-wrap-up
🎯 Goal:
Provide a two-minute closing segment (~250-300 words) summarizing key findings, offering one actionable fact-checking tip, and promoting next week’s show in Leo’s signature charismatic style.
🧠 Initial State:
Pre-loaded Memories:
- 💭 {'kind': 'preference', 'content': 'Encourage listeners to cross-reference clips with metadata tools.', 'importance': 5}
📨 Input Events:
chat_msg
viewer:regular_rita
"Great episode! Any final thoughts before we log off?"
Ready for Testing
Latency by Model (This Suite)
Fastest
- mistralai/mistral-7b-in… 275 ms
- p95 • avg • N 30459 ms • 6444 ms • 7
- [email protected]/Qw… 11478 ms
- p95 • avg • N 14792 ms • 10664 ms • 6
- qwen/qwen-2.5-7b-instru… 24024 ms
- p95 • avg • N 39333 ms • 26733 ms • 12
- qwen/qwen3-14b 26734 ms
- p95 • avg • N 48094 ms • 30590 ms • 12
- qwen/qwen3-8b 28411 ms
- p95 • avg • N 46277 ms • 32324 ms • 10
Slowest
- [email protected]/Qw… 38489 ms
- p95 • avg • N 40633 ms • 38771 ms • 6
- meta-llama/llama-3.1-8b… 28889 ms
- p95 • avg • N 43153 ms • 26967 ms • 11
- qwen/qwen3-8b 28411 ms
- p95 • avg • N 46277 ms • 32324 ms • 10
- qwen/qwen3-14b 26734 ms
- p95 • avg • N 48094 ms • 30590 ms • 12
- qwen/qwen-2.5-7b-instru… 24024 ms
- p95 • avg • N 39333 ms • 26733 ms • 12
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
43130456
Dec. 17, 2025, 12:01 a.m.
59076727
Dec. 16, 2025, 12:01 a.m.
38323419
Dec. 15, 2025, 12:01 a.m.
39975477
Dec. 14, 2025, 12:01 a.m.
38979236
Dec. 13, 2025, 12:01 a.m.
52332051
Dec. 12, 2025, 12:01 a.m.
48384351
Dec. 11, 2025, 12:01 a.m.
40650126
Dec. 10, 2025, 12:01 a.m.
54512095
Dec. 9, 2025, 12:01 a.m.
43267672
Dec. 8, 2025, 12:01 a.m.