Compressed caveman-style prose for AI coding agents — cuts ~65% output tokens while keeping full technical accuracy
96
100%
Does it follow best practices?
Impact
96%
1.00xAverage score across 38 eval scenarios
Passed
No known issues
{
"context": "Tests whether the response correctly explains the GraphQL N+1 problem and provides the standard solution.",
"type": "weighted_checklist",
"checklist": [
{
"name": "Explains why N+1 happens in GraphQL",
"description": "Explains that each field resolver runs independently, so the author resolver fires once per post",
"max_score": 12
},
{
"name": "Recommends DataLoader or batching",
"description": "Suggests DataLoader (or equivalent batching/coalescing library) to batch author lookups into a single query",
"max_score": 15
},
{
"name": "Explains DataLoader mechanism",
"description": "Describes how DataLoader collects IDs during a tick and makes one batched query (WHERE id IN (...))",
"max_score": 10
},
{
"name": "Mentions per-request DataLoader instances",
"description": "Notes that DataLoader instances should be created per-request to avoid stale cache across requests",
"max_score": 8
},
{
"name": "No incorrect information",
"description": "Response does not contain factually wrong statements about GraphQL resolution or DataLoader",
"max_score": 10
}
]
}evals
scenario-1
scenario-2
scenario-3
scenario-4
scenario-5
scenario-6
scenario-7
scenario-8
scenario-9
scenario-10
scenario-11
scenario-12
scenario-13
scenario-14
scenario-15
scenario-16
scenario-17
scenario-18
scenario-19
scenario-20
scenario-21
scenario-22
scenario-23
scenario-24
scenario-25
scenario-26
scenario-27
scenario-28
scenario-29
scenario-30
scenario-31
scenario-32
scenario-33
scenario-34
scenario-35
scenario-36
scenario-37