NC Bench
Overview
About
Tests
Models
Model
Gemini Flash 1.5
Test
Dialogue tags
Scenario
Write 200 words with 10% dialogue
Run ID
2024-07-19-01J35T20WD60G7EJRZE82YYV31
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 10% 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
“You’re sure this is a good idea, Bill?” Molly squinted at the sprawling backyard through the gauzy curtains. Multicolored balloons bobbed in the breeze like drunken jellyfish. "It’s a kids’ party. They’ll have water balloons and pin the tail on the donkey. This whole thing feels… sticky." Bill, clad in a ludicrously oversized clown suit, pulled a face. “You’re saying you’d rather be back at the apartment, watching paint dry?” “No, it’s just… this feels wrong.” “Wrong? Wrong is breaking into a bank and taking five grand. This is kid’s stuff.” Bill patted his vest pocket. “Candy, toys, and enough cash to buy us a trip to the Bahamas. This is a win-win.” He winked. “Besides, what could possibly go wrong?” Molly frowned. “What if someone calls the police?” Her eyes darted to the entrance. “What if they all have cameras?” “Relax. I’ve got this.” Bill gave her a toothy grin. "We’ll be in and out before the birthday cake is even served." He nudged her towards the door. "Come on, let's go."
id
"gen-FMMDpMyIsxhX3E88xOEFU4xYSe8p"
total_cost
0.000216
created_at
"2024-07-19T15:34:08.175544+00:00"
Friday, July 19th, 2024 at 3:34:08 PM GMT+00:00
model
"google/gemini-flash-1.5"
app_id
182717
streamed
false
cancelled
false
provider_name
"Google"
latency
2461
moderation_latency
(null)
generation_time
0
finish_reason
"STOP"
tokens_prompt
68
tokens_completion
257
native_tokens_prompt
69
native_tokens_completion
265
num_media_prompt
(null)
num_media_completion
(null)
origin
"https://ncbench.com/"
usage
0.000216
Evaluation details
Result
Evaluator
Details
Meta Data
1.0360%
Matches word count
n/a
neededClean
false
words
174
0.0000%
Dialogue to Total Word Ratio
Ratio: 67.24%, Deviation: 57.24%
neededClean
false
wordsTotal
174
wordsDialogue
117
0.5180%