NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2024-07-23-01J3G81X38X7EXN8EHRG69KSCV
Run Details
Model:
Llama 3.1 70B
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
3.1s
Cost:
$0.000320
Provider:
Fireworks
Evaluation summary
Score
Evaluator
Details Link
99.38%
Matches word count
(details)
36.79%
Dialogue to Total Word Ratio
(details)
68.0824%
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
"I'm telling you, Molly, this is the score of a lifetime. I mean, who needs Fort Knox when you've got a room full of sugared-up kids and their presents?" "Bill, are you insane? This is a children's party, not a high-stakes heist. What's the plan, exactly? Sneak in disguised as a clown and make off with the loot?" "Hey, I've got the perfect clown costume. I've been practicing my squirt-in-the-ear trick for weeks. But no, seriously, we create a diversion, grab the gifts, and make a quick exit. It's a cakewalk." "A cakewalk? Bill, we're talking about a room full of kids on a sugar high. It's going to be a war zone. What if we get caught?" "What's the worst that could happen? We get booed out of the party and have to return all the presents? I mean, come on, Molly. It's worth the risk for the haul we'll get." "Haul? You mean a bunch of Pokémon toys and Barbie dolls?" "Hey, don't knock it. Those things are worth a pretty penny on the black market. Plus, think of the satisfaction we'll get from ruining a bunch of kids' birthday parties." "Oh, I can already feel the moral fiber of my being disintegrating."
id
"gen-y9cXdUA6wIXTOI7TxKmiPSkmst44"
total_cost
0.0003204
upstream_id
"8f7b0815-38d8-4b39-8a14-0d4fd7e9e622"
created_at
"2024-07-23T16:51:09.258019+00:00"
7/23/2024, 4:51:09 PM
model
"meta-llama/llama-3.1-70b-instruct"
app_id
182717
streamed
false
cancelled
false
provider_name
"Fireworks"
latency
2893
moderation_latency
(null)
generation_time
0
finish_reason
"stop"
tokens_prompt
102
tokens_completion
284
native_tokens_prompt
70
native_tokens_completion
286
num_media_prompt
(null)
num_media_completion
(null)
origin
"https://ncbench.com/"
usage
0.0003204
Evaluation details
Result
Evaluator
Details
Meta Data
99.38%
Matches word count
n/a
neededClean
false
words
205
36.79%
Dialogue to Total Word Ratio
Ratio: 100.00%, Deviation: 10.00%
neededClean
false
wordsTotal
210
wordsDialogue
210
68.0824%