NC Bench
Overview
About
Tests
Models
Model
Llama 3.2 11B (Vision)
Test
Dialogue tags
Scenario
Write 200 words with 50% dialogue
Run ID
2024-09-26-01J8PYW9PWRNN5YDFANKCGQTCS
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 50% words of dialogue between two thieves named Molly and Bill who are planning to break into a children's party to steal all the presents. Only include the prose of the scene without any markdown or commentary.
Model response
Messages
Meta
AI
The sun-drenched parking lot was a typical Sunday afternoon crowd, strollers and screaming kids the norm. Not exactly the most thrilling hideout, but Bill had rigged a fake ventilation unit to create a convincing cloud of dust, keeping prying eyes at bay. Molly leaned against the wall, blowing on a lukewarm coffee, studying the blueprints of the party venue. "I swear, Bill, this is the sixth time I've read this. I think I've got it memorized. Two guards, party favors for the masses, and a piƱata that's just asking to be dropped on someone's head." Bill scrubbed his jaw, eyes squinting. "You know as well as I do that a really good heist starts with the planning. Details, Molly, details! What's the MO here?" "Easy in, wave hello to Santa, dive for the presents when the room clears." "Not bad, not bad," Bill said, pocketing his pen. "But what about contingencies? What if the cake-cutter gets an emergency beeper-on-rent call and brings in reinforcements?" Molly snorted. "The cake-cutter doesn't even wear a real uniform. He's more worried about spilled orange juice stains than getting caught." "Well, a 'real' party would have a secret password, or a surprise performer." "Surprise performer?" Molly mouthed it like it was the worst idea she'd ever heard. "Santa in a wrestling singlet? Help me, Bill."
id
"gen-lD0TwxEyQsOMcskReYsQYGG5evFn"
total_cost
0.000076
upstream_id
"536a954b-301a-45c6-8789-dd1a03796f75"
created_at
"2024-09-26T10:43:58.213325+00:00"
Thursday, September 26th, 2024 at 10:43:58 AM GMT+00:00
model
"meta-llama/llama-3.2-11b-vision-instruct"
app_id
182717
streamed
true
cancelled
false
provider_name
"Fireworks"
latency
146
moderation_latency
(null)
generation_time
2556
finish_reason
"stop"
tokens_prompt
68
tokens_completion
308
native_tokens_prompt
71
native_tokens_completion
309
native_tokens_reasoning
(null)
num_media_prompt
(null)
num_media_completion
(null)
origin
"https://ncbench.com/"
usage
0.000076
Evaluation details
Result
Evaluator
Details
Meta Data
2.0116%
Matches word count
n/a
neededClean
false
words
225
10.7365%
Dialogue to Total Word Ratio
Ratio: 62.22%, Deviation: 12.22%
neededClean
false
wordsTotal
225
wordsDialogue
140
6.3740%