NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2024-07-19-01J35SZE2FM0E0J4XYGV01W87X
Run Details
Model:
DeepSeek-V2 Chat
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
19.0s
Cost:
$0.000107
Provider:
DeepSeek
Evaluation summary
Score
Evaluator
Details Link
68.10%
Matches word count
(details)
97.25%
Dialogue to Total Word Ratio
(details)
82.6767%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 90% 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
Molly: "Alright, Bill, here’s the plan. We dress up as clowns, blend in, and swipe all the presents when no one’s looking." Bill: "Clowns? Really? You do know I’m colorblind, right? How am I supposed to match my polka dots?" Molly: "It’s a children’s party, Bill. They won’t care about your fashion sense. Just slap on some makeup and a wig. We’ll be in and out before they notice the cake’s missing." Bill: "And what if they ask us to juggle or something? I can’t juggle." Molly: "You can’t juggle? What kind of thief are you? Just pretend to drop the balls a lot. It’s part of the act." Bill: "And what’s my clown name? Should I go with something tough, like ‘Killer’?" Molly: "‘Killer’? At a children’s party? You’re going with ‘Happy’. It’s less terrifying." Bill: "‘Happy’? I look more like a ‘Dopey’ to be honest." Molly: "Just stick to the plan, ‘Dopey’. We’ll be rich by bedtime." Bill: "Rich with toys. Great. Can’t wait to explain that to the cops." Molly: "Oh, come on, Bill. It’s a children’s party. What could possibly go wrong?"
id
"gen-vsCBDayvj2jQisbw6nIOSrM3dIQZ"
total_cost
0.00010654
upstream_id
"e23a8425b9c21df0814aa86bba6441a0"
created_at
"2024-07-19T15:32:59.840728+00:00"
7/19/2024, 3:32:59 PM
model
"deepseek/deepseek-chat"
app_id
182717
streamed
false
cancelled
false
provider_name
"DeepSeek"
latency
18911
moderation_latency
(null)
generation_time
0
finish_reason
"stop"
tokens_prompt
68
tokens_completion
298
native_tokens_prompt
73
native_tokens_completion
344
num_media_prompt
(null)
num_media_completion
(null)
origin
"https://ncbench.com/"
usage
0.00010654
Evaluation details
Result
Evaluator
Details
Meta Data
68.10%
Matches word count
n/a
neededClean
false
words
186
97.25%
Dialogue to Total Word Ratio
Ratio: 94.09%, Deviation: 4.09%
neededClean
false
wordsTotal
186
wordsDialogue
175
82.6767%