NC Bench
Overview
About
Tests
Models
Model
Phi-3.5 Mini 128k
Test
Dialogue tags
Scenario
Write 200 words with 90% dialogue
Run ID
2024-08-26-01J68D3NB688W5AY0PR3SMY8MV
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: Okay, Bill, remember, stealth is our main code. Kids' parties are chaotic, but it's perfect—piles of presents! Bill: Right. We hit at twelve. The parents are still blissfully unaware. I'll play Santa. Molly: Got it. I'll be a clown—a floating, deflating clown. Non-threatening. Bill: Ha, classic. Now, if we just distract them with balloons… Molly: Swoon. I'll make them think they're in a circus. But hey, what about the dog? It comes out at parties. Bill: Deploy, Sparky. You distract with fetch, I'll time my sneak out. Molly: Snickers, two dogs won't budge. But the kids will be entranced. Bill: And there's the exit strategy. The backyard door's a sliding glass pane. Molly: Slide glass, slide. Remember, we've slept on mirror shards. Bill: No ice this time, my clown friend. Bye, Molly. Molly: Bye. This is going to be legendary—clown and Sparky stolen presents. Huzzah!" (Fade out as they begin their comical and cheeky plan to plunder the party with laughter and floppy shoes.)
id
"gen-Bvr0alSgoM3ZXPc8esG8AHtsx5qd"
total_cost
0.0000442
upstream_id
"chat-229d88741d7b4cd49b0aa95e105d7be3"
created_at
"2024-08-26T22:32:45.525999+00:00"
Monday, August 26th, 2024 at 10:32:45 PM GMT+00:00
model
"microsoft/phi-3.5-mini-128k-instruct"
app_id
182717
streamed
true
cancelled
false
provider_name
"Azure"
latency
127
moderation_latency
(null)
generation_time
8670
finish_reason
"stop"
tokens_prompt
82
tokens_completion
261
native_tokens_prompt
84
native_tokens_completion
358
num_media_prompt
(null)
num_media_completion
(null)
origin
"https://ncbench.com/"
usage
0.0000442
Evaluation details
Result
Evaluator
Details
Meta Data
0.0098%
Matches word count
n/a
neededClean
false
words
169
99.9762%
Dialogue to Total Word Ratio
Ratio: 88.76%, Deviation: 1.24%
neededClean
false
wordsTotal
169
wordsDialogue
150
49.9930%