lutzroeder/netron

11.5

Adjusted Score

11.5

Raw Score

100%

Time Factor

2026-07-13

Last Push

33.2K

Stars

JavaScript

Language

265.3K

Lines of Code

215

Files

349

Pattern Hits

2026-07-14

Scan Date

1.40

HC Hit Rate

What These Metrics Mean

Adjusted Score: Primary synthetic code indicator. Raw score normalised per 1,000 lines of code and multiplied by the temporal discount factor. This is the definitive comparative metric — use it to rank repositories by AI authorship density.
Raw Score: The unmodified sum of all severity-weighted, context-multiplied pattern match scores before temporal discounting. Reflects the absolute signal strength independent of when the repository was last active.
Time Factor: The temporal discount multiplier (0–100%) applied to the raw score. Repositories last updated before ChatGPT's launch (Nov 2022) receive a 5% factor. Full signal is only assigned to repositories active in the post-adoption era (Jan 2024+).
Pattern Hits: Total count of individual pattern matches across all files and categories. A high hit count with a low score may indicate a very large codebase with isolated AI snippets; a low count with a high score indicates dense, concentrated AI signatures.
HC Hit Rate: High+Critical pattern hits per file, averaged across the repository. This orthogonal signal catches repositories where a few files are densely packed with high-severity AI tells — a strong indicator even when the normalised score appears moderate due to codebase size.
Lines of Code / Files: Total lines and files analysed. The scanner examines 94 file extensions. These denominators are used to normalise the score, enabling fair comparison between repositories of vastly different sizes.

Score History

This chart maps the temporal evolution of the adjusted synthetic code score across successive scan runs. An upward trajectory indicates ongoing incorporation of AI-generated code or expanding LLM-assisted scaffolding; a stable or declining trajectory may reflect active human refactoring, code removal, or the adoption of stricter authorship policies. The dashed secondary line (right axis) independently tracks total raw pattern hit count, which can diverge from the normalised score when codebase size changes significantly between scans.

Severity Breakdown

Classifies detected patterns by their diagnostic confidence and structural impact. CRITICAL patterns (coefficient 10) represent definitive synthetic signatures — hallucinated imports, explicit LLM attribution metadata — virtually never produced by human authors. HIGH (5) indicates strong structural tells such as cross-file repetition or cross-linguistic idioms. MEDIUM (2) covers recognisable conversational padding and AI-specific vocabulary. LOW (1) captures subtle indicators like tautological comments and generic boilerplate that require density to carry independent signal.

CRITICAL 300HIGH 0MEDIUM 0LOW 49

Directory Score Breakdown

This horizontal bar chart decomposes the repository's raw synthetic code score by top-level directory, allowing you to pinpoint precisely which modules or components carry the highest AI authorship density. Directories with disproportionately high scores relative to their size warrant targeted manual review: concentrated AI signatures often trace back to mass-generated configuration layers, auto-ported test suites, LLM-scaffolded boilerplate classes, or entire subsystems authored under heavy copilot assistance. Use this view to prioritise your human code-review effort.

Pattern Findings

The scanner identified 349 distinct pattern matches across 10 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a statistically significant AI authorship indicator. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.

Reading the findings table: The Severity column indicates the diagnostic confidence level (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — comment-context matches receive a ×1.5 weight because LLMs systematically over-annotate. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier as dense clusters constitute far stronger evidence of synthetic authorship than isolated hits.

Hallucination Indicators300 hits · 3002 pts

Severity	File	Line	Snippet	Context
CRITICAL⚡	source/pytorch-schema.js	284	case 1: return torch.jit.mobile.serialization.Int.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	285	case 2: return torch.jit.mobile.serialization.Bool.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	286	case 3: return torch.jit.mobile.serialization.Double.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	287	case 4: return torch.jit.mobile.serialization.ComplexDouble.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	288	case 5: return torch.jit.mobile.serialization.TensorMetadata.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	289	case 6: return torch.jit.mobile.serialization.String.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	290	case 7: return torch.jit.mobile.serialization.List.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	291	case 8: return torch.jit.mobile.serialization.Tuple.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	292	case 9: return torch.jit.mobile.serialization.Dict.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	293	case 10: return torch.jit.mobile.serialization.Object.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	294	case 11: return torch.jit.mobile.serialization.IntList.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	295	case 12: return torch.jit.mobile.serialization.DoubleList.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	296	case 13: return torch.jit.mobile.serialization.BoolList.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	297	case 14: return torch.jit.mobile.serialization.Device.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	298	case 15: return torch.jit.mobile.serialization.EnumValue.decode(reader, position);	CODE
CRITICAL⚡	source/pytorch-schema.js	299	case 16: return torch.jit.mobile.serialization.Function.decode(reader, position);	CODE
CRITICAL	source/pytorch-schema.js	331	return torch.jit.mobile.serialization.Module.decode(reader, reader.root);	CODE
CRITICAL⚡	source/megengine-schema.js	2192	case 1: return mgb.serialization.fbs.v2.DefaultTensorFormat.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2193	case 2: return mgb.serialization.fbs.v2.Image2DPackedTensorFormat.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2194	case 3: return mgb.serialization.fbs.v2.LowbitsAlignedTensorFormat.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2243	case 1: return mgb.serialization.fbs.param.Empty.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2244	case 2: return mgb.serialization.fbs.param.Axis.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2245	case 3: return mgb.serialization.fbs.param.Convolution.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2246	case 4: return mgb.serialization.fbs.param.MaskPropagate.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2247	case 5: return mgb.serialization.fbs.param.ConvPooling.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2248	case 6: return mgb.serialization.fbs.param.ConvBias.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2249	case 7: return mgb.serialization.fbs.param.SeparableConv.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2250	case 8: return mgb.serialization.fbs.param.Images2Neibs.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2251	case 9: return mgb.serialization.fbs.param.Pooling.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2252	case 10: return mgb.serialization.fbs.param.LRN.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2253	case 11: return mgb.serialization.fbs.param.BN.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2254	case 12: return mgb.serialization.fbs.param.ROIPooling.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2255	case 13: return mgb.serialization.fbs.param.WarpPerspective.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2256	case 14: return mgb.serialization.fbs.param.SpatialTfGridGenerator.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2257	case 15: return mgb.serialization.fbs.param.SpatialTfSampler.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2258	case 16: return mgb.serialization.fbs.param.MGBAddUpdate.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2259	case 17: return mgb.serialization.fbs.param.Elemwise.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2260	case 18: return mgb.serialization.fbs.param.ElemwiseMultiType.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2261	case 19: return mgb.serialization.fbs.param.PowC.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2262	case 20: return mgb.serialization.fbs.param.MatrixMul.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2263	case 21: return mgb.serialization.fbs.v2.DeprecatedParam.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2264	case 22: return mgb.serialization.fbs.param.SVD.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2265	case 23: return mgb.serialization.fbs.param.Reduce.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2266	case 24: return mgb.serialization.fbs.param.Cumsum.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2267	case 25: return mgb.serialization.fbs.param.CondTake.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2268	case 26: return mgb.serialization.fbs.param.Argsort.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2269	case 27: return mgb.serialization.fbs.param.IndexingRemap.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2270	case 28: return mgb.serialization.fbs.param.MGBSleep.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2271	case 29: return mgb.serialization.fbs.param.Linspace.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2272	case 30: return mgb.serialization.fbs.param.LinspaceFull.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2273	case 31: return mgb.serialization.fbs.param.Eye.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2274	case 32: return mgb.serialization.fbs.param.UniformRNG.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2275	case 33: return mgb.serialization.fbs.param.GaussianRNG.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2276	case 34: return mgb.serialization.fbs.param.Flip.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2277	case 35: return mgb.serialization.fbs.param.Rotate.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2278	case 36: return mgb.serialization.fbs.param.ROICopy.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2279	case 37: return mgb.serialization.fbs.param.CvtColor.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2280	case 38: return mgb.serialization.fbs.param.WarpAffine.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2281	case 39: return mgb.serialization.fbs.param.GaussianBlur.decode(reader, position);	CODE
CRITICAL⚡	source/megengine-schema.js	2282	case 40: return mgb.serialization.fbs.param.Resize.decode(reader, position);	CODE
240 more matches not shown…

Deep Nesting12 hits · 12 pts

Severity	File	Line	Context
LOW	tools/keras_script.py	97	CODE
LOW	tools/keras_script.py	149	CODE
LOW	tools/keras_script.py	174	CODE
LOW	tools/nnabla_script.py	19	CODE
LOW	tools/nnabla_script.py	102	CODE
LOW	tools/nnabla_script.py	20	CODE
LOW	source/server.py	206	CODE
LOW	source/server.py	67	CODE
LOW	source/onnx.py	97	CODE
LOW	source/pytorch.py	58	CODE
LOW	source/pytorch.py	80	CODE
LOW	source/pytorch.py	148	CODE

Fake / Example Data10 hits · 10 pts

Severity	File	Line	Snippet	Context
LOW	source/tf-proto.js	1438	tensorflow.AttrValue.valueSet = tensorflow.AttrValue.valueSet \|\| new Set(["s", "i", "f", "b", "type", "shape", "	CODE
LOW	source/tf-proto.js	1519	case "placeholder":	CODE
LOW	source/tf-proto.js	1559	if ('placeholder' in obj) {	CODE
LOW	source/tf.js	977	case 'placeholder': {	CODE
LOW	source/tf.js	978	type = 'placeholder';	CODE
LOW	source/cntk.js	179	// VariableKind { 0: 'input', 1: 'output', 2: 'parameter', 3: 'constant', 4: 'placeholder' }	COMMENT
LOW	source/pytorch.js	235	if (obj.op === 'placeholder') {	CODE
LOW	source/pytorch.js	664	} else if (obj.op === 'placeholder') {	CODE
LOW	source/pytorch.js	1076	if (obj.op === 'placeholder') {	CODE
LOW	source/view.js	5115	this._query.setAttribute('placeholder', 'Search');	CODE

Over-Commented Block10 hits · 10 pts

Severity	File	Line	Snippet	Context
LOW	test/backend.py	81		COMMENT
LOW	source/gguf.js	921	// Resolve display type/category for a component group from metadata	COMMENT
LOW	source/dagre.js	161	// Returns the amount of slack for the given edge.	COMMENT
LOW	source/dagre.js	181	// Constructs a spanning tree with tight edges and adjusted the input node's ranks to achieve this.	COMMENT
LOW	source/dagre.js	241	// This algorithm scales well and is fast in practice, it yields rather poor solutions.	COMMENT
LOW	source/dagre.js	281	}	COMMENT
LOW	source/dagre.js	301	// edges in the tree have no slack (difference between length of edge	COMMENT
LOW	source/dagre.js	881	queue.push(v);	COMMENT
LOW	source/dagre.js	1241	// This makes it easy to walk the hierarchy of movable nodes during ordering.	COMMENT
LOW	source/dagre.js	1801	};	COMMENT

Modern Structural Boilerplate5 hits · 5 pts

Severity	File	Line	Snippet	Context
LOW	tools/pytorch_script.py	23	logger = logging.getLogger(__name__)	CODE
LOW	test/backend.py	17	logger = logging.getLogger(__name__)	CODE
LOW	test/measures.py	11	logger = logging.getLogger(__name__)	CODE
LOW	source/server.py	21	logger = logging.getLogger(__name__)	CODE
LOW	source/__init__.py	10	__all__ = ["start", "stop", "status", "wait", "serve", "widget", "__version__"]	CODE

Unused Imports4 hits · 4 pts

Severity	File	Line	Context
LOW	source/__init__.py	8	CODE
LOW	source/__init__.py	8	CODE
LOW	source/__init__.py	8	CODE
LOW	source/__init__.py	8	CODE

Verbosity Indicators2 hits · 3 pts

Severity	File	Line	Snippet	Context
LOW	tools/tablegen.js	1655	// Step 1: Build initial template bindings for this record from its immediate parents	COMMENT
LOW	tools/tablegen.js	1681	// Step 2: Process parents and flatten fields	COMMENT

Structural Annotation Overuse2 hits · 3 pts

Severity	File	Line	Snippet	Context
LOW	tools/tablegen.js	1655	// Step 1: Build initial template bindings for this record from its immediate parents	COMMENT
LOW	tools/tablegen.js	1681	// Step 2: Process parents and flatten fields	COMMENT

Hyper-Verbose Identifiers3 hits · 3 pts

Severity	File	Line	Snippet	Context
LOW	test/backend.py	52	def _test_torchscript_transformer():	CODE
LOW	test/backend.py	71	def _test_torchscript_quantized():	CODE
LOW	source/server.py	46	class _HTTPRequestHandler(http.server.BaseHTTPRequestHandler):	CODE

AI Structural Patterns1 hit · 1 pts

Severity	File	Line	Snippet	Context
LOW	source/server.py	160		CODE

Analysis Overview

What These Metrics Mean

Score History

Severity Breakdown

Directory Score Breakdown

Pattern Findings