Deng-m1andCursor
232fbc6dd8
TypeScript library + CLI implementing single-token behavioral fingerprinting of LLMs behind OpenAI-compatible APIs, per 'One Token Is Enough' (Bruckner, arXiv:2607.10252). - Zero runtime dependencies (Node >= 18.17 built-in fetch) - fingerprint / compare / verify API + llm-fingerprint CLI - Jensen-Shannon divergence over per-cell answer distributions - Reasoning-model aware (auto-disables hidden thinking channels) - Bundled sample references for 11 popular models (CC-BY-4.0, derived from the paper's public Zenodo dataset) - Unit + integration tests via node --test Co-authored-by: Cursor <[email protected]>
24 lines
570 B
JSON
24 lines
570 B
JSON
{
|
|
"compilerOptions": {
|
|
"target": "ES2022",
|
|
"module": "NodeNext",
|
|
"moduleResolution": "NodeNext",
|
|
"lib": ["ES2022"],
|
|
"outDir": "dist",
|
|
"rootDir": "src",
|
|
"declaration": true,
|
|
"declarationMap": true,
|
|
"sourceMap": true,
|
|
"strict": true,
|
|
"noUncheckedIndexedAccess": false,
|
|
"noImplicitOverride": true,
|
|
"noFallthroughCasesInSwitch": true,
|
|
"forceConsistentCasingInFileNames": true,
|
|
"isolatedModules": true,
|
|
"verbatimModuleSyntax": true,
|
|
"skipLibCheck": true,
|
|
"types": ["node"]
|
|
},
|
|
"include": ["src"]
|
|
}
|