Andyyyy64/whichllm

28.7

Adjusted Score

28.7

Raw Score

100%

Time Factor

2026-07-08

Last Push

5.8K

Stars

Python

Language

21.2K

Lines of Code

118

Files

558

Pattern Hits

2026-07-14

Scan Date

0.05

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 0HIGH 6MEDIUM 3LOW 549

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 558 distinct pattern matches across 13 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.

Hyper-Verbose Identifiers324 hits · 338 pts

Severity	File	Line	Snippet	Context
LOW⚡	tests/test_utils.py	10	def test_cache_dir_defaults_to_dot_cache(monkeypatch):	CODE
LOW⚡	tests/test_utils.py	16	def test_cache_dir_respects_xdg_cache_home(monkeypatch):	CODE
LOW⚡	tests/test_utils.py	22	def test_cache_dir_falls_back_on_empty_xdg(monkeypatch):	CODE
LOW⚡	tests/test_utils.py	28	def test_cache_dir_ignores_relative_xdg(monkeypatch):	CODE
LOW⚡	tests/test_utils.py	52	def test_parse_fractional_suffix():	CODE
LOW⚡	tests/test_utils.py	57	def test_parse_whitespace_is_stripped():	CODE
LOW⚡	tests/test_utils.py	61	def test_parse_rejects_invalid_string():	CODE
LOW⚡	tests/test_utils.py	71	def test_parse_rejects_negative():	CODE
LOW⚡	tests/test_utils.py	76	def test_click_type_passes_int_through():	CODE
LOW⚡	tests/test_utils.py	80	def test_click_type_parses_shorthand():	CODE
LOW	tests/test_output_formatting.py	25	def test_format_speed_colors_by_runtime_speed_not_confidence():	CODE
LOW	tests/test_gtx1650_variants.py	32	def test_variant_table_present_and_reexported():	CODE
LOW	tests/test_gtx1650_variants.py	54	def test_unknown_clock_falls_back_to_curated_default():	CODE
LOW	tests/test_gtx1650_variants.py	68	def test_non_variant_card_ignores_memory_clock():	CODE
LOW	tests/test_gtx1650_variants.py	147	def test_smi_na_clock_falls_back_to_curated(monkeypatch):	CODE
LOW	tests/test_gtx1650_variants.py	166	def test_smi_both_queries_fail_returns_empty(monkeypatch):	CODE
LOW	tests/test_compatibility.py	73	def test_usable_vram_budget_can_turn_full_gpu_into_partial_offload():	CODE
LOW	tests/test_compatibility.py	86	def test_ram_budget_limits_partial_offload_pool():	CODE
LOW	tests/test_compatibility.py	98	def test_ram_budget_caps_shared_memory_gpu_fit_pool():	CODE
LOW	tests/test_compatibility.py	126	def test_shared_memory_manual_vram_override_caps_available_gpu_memory():	CODE
LOW	tests/test_compatibility.py	153	def test_shared_memory_amd_apu_uses_system_memory_pool():	CODE
LOW	tests/test_compatibility.py	182	def test_windows_shared_memory_amd_apu_does_not_emit_rocm_warning():	CODE
LOW	tests/test_compatibility.py	211	def test_shared_memory_igpu_is_not_summed_with_dedicated_gpu():	CODE
LOW	tests/test_compatibility.py	244	def test_homogeneous_multi_gpu_uses_conservative_fit_budget():	CODE
LOW	tests/test_compatibility.py	282	def test_heterogeneous_multi_gpu_warns_about_split_assumptions():	CODE
LOW	tests/test_compatibility.py	318	def test_multiple_shared_memory_gpus_are_not_summed():	CODE
LOW	tests/test_compatibility.py	352	def test_apple_silicon_does_not_double_count_unified_memory():	CODE
LOW	tests/test_compatibility.py	385	def test_apple_silicon_full_gpu_fit():	CODE
LOW	tests/test_compatibility.py	413	def test_apple_silicon_vendor_guard_handles_legacy_shared_memory_false():	CODE
LOW	tests/test_compatibility.py	461	def test_low_compute_capability():	CODE
LOW⚡	tests/test_compatibility.py	479	def test_context_fits_true_when_model_supports():	CODE
LOW⚡	tests/test_compatibility.py	487	def test_context_fits_false_when_model_too_small():	CODE
LOW⚡	tests/test_compatibility.py	496	def test_context_fits_unknown_is_true():	CODE
LOW	tests/test_ranker.py	38	def test_ranker_picks_highest_scoring_variant():	CODE
LOW	tests/test_ranker.py	74	def test_quant_filter_applies_to_non_gguf_models():	CODE
LOW	tests/test_ranker.py	131	def test_darwin_backend_filters_out_non_gguf_models():	CODE
LOW	tests/test_ranker.py	163	def test_cpu_only_backend_filters_out_non_gguf_models():	CODE
LOW	tests/test_ranker.py	211	def test_rank_models_clamps_non_positive_top_n():	CODE
LOW	tests/test_ranker.py	232	def test_popularity_has_no_effect_with_direct_benchmark():	CODE
LOW	tests/test_ranker.py	277	def test_general_profile_excludes_specialized_models():	CODE
LOW	tests/test_ranker.py	323	def test_require_direct_top_prioritizes_direct_benchmark():	CODE
LOW	tests/test_ranker.py	371	def test_min_params_filter_excludes_small_models():	CODE
LOW	tests/test_ranker.py	418	def test_general_profile_prefers_full_gpu_when_direct_is_partial():	CODE
LOW	tests/test_ranker.py	454	def test_family_dedup_prefers_direct_when_enabled():	CODE
LOW	tests/test_ranker.py	494	def test_full_gpu_estimated_ranks_above_partial_direct():	CODE
LOW	tests/test_ranker.py	535	def test_strong_partial_offload_not_buried_below_weaker_full_gpu():	CODE
LOW	tests/test_ranker.py	638	def test_moe_partial_offload_penalty_uses_active_working_set():	CODE
LOW	tests/test_ranker.py	658	def test_evidence_strict_filters_out_estimated_models():	CODE
LOW	tests/test_ranker.py	706	def test_evidence_base_keeps_base_model_match_and_drops_line_interp():	CODE
LOW	tests/test_ranker.py	758	def test_unknown_speed_heavy_partial_offload_does_not_top_rank():	CODE
LOW	tests/test_ranker.py	827	def test_fit_filter_full_gpu_excludes_partial_offload_and_cpu_only():	CODE
LOW	tests/test_ranker.py	887	def test_fit_filter_full_gpu_returns_empty_when_no_full_gpu_candidate():	CODE
LOW	tests/test_ranker.py	916	def test_multi_gpu_speed_confidence_is_low():	CODE
LOW	tests/test_ranker.py	980	def test_benchmark_source_and_confidence_exposed_for_direct():	CODE
LOW	tests/test_ranker.py	1010	def test_benchmark_source_and_confidence_exposed_for_estimated():	CODE
LOW	tests/test_ranker.py	1040	def test_benchmark_source_and_confidence_exposed_for_self_reported():	CODE
LOW	tests/test_ranker.py	1071	def test_benchmark_source_and_confidence_exposed_for_none():	CODE
LOW	tests/test_ranker.py	1101	def test_ctx_penalty_demotes_non_fitting():	CODE
LOW	tests/test_p1_p3_regressions.py	407	def test_prequantized_repo_skips_synthesis():	CODE
LOW	tests/test_p1_p3_regressions.py	427	def test_newer_generation_beats_older_at_same_size():	CODE
264 more matches not shown…

Unused Imports147 hits · 140 pts

Severity	File	Line	Context
LOW	tests/test_p1_p3_regressions.py	18	CODE
LOW	tests/test_intel_gpu.py	3	CODE
LOW	tests/test_cache_encoding.py	3	CODE
LOW	tests/test_asahi_detection.py	3	CODE
LOW	tests/test_aa_index.py	9	CODE
LOW	tests/test_windows_gpu.py	3	CODE
LOW	tests/test_r3_regressions.py	18	CODE
LOW	tests/test_amd_detection.py	3	CODE
LOW	tests/test_windows_cpu_detection.py	3	CODE
LOW	scripts/import_livebench_csv.py	11	CODE
LOW	src/whichllm/constants.py	8	CODE
LOW	src/whichllm/constants.py	8	CODE
LOW	src/whichllm/constants.py	8	CODE
LOW	src/whichllm/constants.py	13	CODE
LOW	src/whichllm/constants.py	13	CODE
LOW	src/whichllm/constants.py	13	CODE
LOW	src/whichllm/constants.py	13	CODE
LOW	src/whichllm/constants.py	13	CODE
LOW	src/whichllm/constants.py	13	CODE
LOW	src/whichllm/constants.py	13	CODE
LOW	src/whichllm/constants.py	13	CODE
LOW	src/whichllm/constants.py	13	CODE
LOW	src/whichllm/constants.py	24	CODE
LOW	src/whichllm/constants.py	24	CODE
LOW	src/whichllm/constants.py	24	CODE
LOW	src/whichllm/constants.py	29	CODE
LOW	src/whichllm/constants.py	29	CODE
LOW	src/whichllm/constants.py	29	CODE
LOW	src/whichllm/cli.py	3	CODE
LOW	src/whichllm/utils.py	1	CODE
LOW	src/whichllm/hardware/detector.py	3	CODE
LOW	src/whichllm/hardware/gpu_db.py	27	CODE
LOW	src/whichllm/hardware/intel.py	3	CODE
LOW	src/whichllm/hardware/nvidia.py	3	CODE
LOW	src/whichllm/hardware/memory.py	3	CODE
LOW	src/whichllm/hardware/cpu.py	3	CODE
LOW	src/whichllm/hardware/gpu_simulator.py	7	CODE
LOW	src/whichllm/hardware/types.py	1	CODE
LOW	src/whichllm/hardware/windows.py	3	CODE
LOW	src/whichllm/hardware/apple.py	3	CODE
LOW	src/whichllm/hardware/amd.py	3	CODE
LOW	src/whichllm/output/plan.py	3	CODE
LOW	src/whichllm/output/formatting.py	3	CODE
LOW	src/whichllm/output/display.py	19	CODE
LOW	src/whichllm/output/display.py	20	CODE
LOW	src/whichllm/output/display.py	20	CODE
LOW	src/whichllm/output/display.py	20	CODE
LOW	src/whichllm/output/display.py	25	CODE
LOW	src/whichllm/output/display.py	26	CODE
LOW	src/whichllm/output/display.py	27	CODE
LOW	src/whichllm/output/display.py	27	CODE
LOW	src/whichllm/output/display.py	28	CODE
LOW	src/whichllm/output/upgrade.py	3	CODE
LOW	src/whichllm/output/json_output.py	3	CODE
LOW	src/whichllm/output/markdown.py	3	CODE
LOW	src/whichllm/output/ranking.py	3	CODE
LOW	src/whichllm/models/benchmark_lineage.py	3	CODE
LOW	src/whichllm/models/gguf.py	3	CODE
LOW	src/whichllm/models/sliding_window.py	3	CODE
LOW	src/whichllm/models/benchmark_cache.py	3	CODE
87 more matches not shown…

Cross-Language Confusion5 hits · 30 pts

Severity	File	Line	Snippet	Context
HIGH	tests/test_aa_index.py	32	fragment inside ``self.__next_f.push([n, "..."])``."""	STRING
HIGH⚡	tests/test_aa_index.py	44	"<script>self.__next_f.push([0])</script>"	CODE
HIGH⚡	tests/test_aa_index.py	45	f"<script>self.__next_f.push([1,{chunk}])</script>"	CODE
HIGH	src/whichllm/hardware/windows.py	112	"$dedicated = $null; "	CODE
HIGH	src/whichllm/models/benchmark_sources/aa_index.py	215	# App Router streaming format: model data arrives in ``self.__next_f.push([n,	COMMENT

Modern Structural Boilerplate22 hits · 22 pts

Severity	File	Line	Snippet	Context
LOW	tests/test_asahi_detection.py	151	def _setup_asahi_sysfs(tmp_path: Path) -> None:	CODE
LOW	src/whichllm/constants.py	35	__all__ = [	CODE
LOW	src/whichllm/hardware/detector.py	17	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/hardware/gpu_db.py	35	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/hardware/intel.py	17	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/hardware/nvidia.py	13	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/hardware/cpu.py	11	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/hardware/gpu_simulator.py	26	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/hardware/windows.py	14	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/hardware/apple.py	14	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/hardware/amd.py	15	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/output/display.py	30	__all__ = [	CODE
LOW	src/whichllm/models/benchmark_cache.py	11	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/models/benchmark.py	65	__all__ = [	CODE
LOW	src/whichllm/models/benchmark_fetch.py	14	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/models/cache.py	11	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/models/fetcher.py	78	__all__ = [	CODE
LOW	src/whichllm/models/hf.py	15	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/models/benchmark_sources/aa_index.py	27	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/models/benchmark_sources/vision.py	25	logger = logging.getLogger(__name__)	CODE
LOW	src/whichllm/models/benchmark_sources/__init__.py	34	__all__ = [	CODE
LOW	src/whichllm/models/benchmark_sources/aider.py	22	logger = logging.getLogger(__name__)	CODE

Excessive Try-Catch Wrapping19 hits · 22 pts

Severity	File	Line	Snippet	Context
LOW⚡	src/whichllm/cli.py	635	except Exception as e:	CODE
LOW⚡	src/whichllm/cli.py	649	except Exception as e:	CODE
LOW	src/whichllm/cli.py	715	except Exception as e:	CODE
LOW	src/whichllm/cli.py	790	except Exception as e:	CODE
LOW	src/whichllm/cli.py	876	except Exception as e:	CODE
LOW	src/whichllm/cli.py	888	except Exception:	CODE
LOW	src/whichllm/cli.py	965	except Exception as e:	CODE
LOW	src/whichllm/hardware/gpu_db.py	129	except Exception as exc: # pragma: no cover - dbgpu is a hard dependency	CODE
LOW	src/whichllm/hardware/nvidia.py	152	except Exception:	CODE
LOW	src/whichllm/hardware/nvidia.py	176	except Exception as clock_err:	CODE
LOW	src/whichllm/hardware/nvidia.py	191	except Exception:	CODE
LOW	src/whichllm/hardware/cpu.py	129	except Exception as e:	CODE
LOW	src/whichllm/hardware/cpu.py	149	except Exception:	CODE
LOW	src/whichllm/hardware/cpu.py	185	except Exception:	CODE
LOW	src/whichllm/hardware/cpu.py	202	except Exception:	CODE
LOW	src/whichllm/hardware/cpu.py	212	except Exception:	CODE
MEDIUM	src/whichllm/hardware/cpu.py	54	def _cpu_name_from_wmic() -> str \| None:	CODE
LOW	src/whichllm/hardware/amd.py	259	except Exception:	CODE
MEDIUM	src/whichllm/hardware/amd.py	73	def _detect_from_lspci() -> list[str]:	CODE

Deep Nesting21 hits · 20 pts

Severity	File	Line	Context
LOW	src/whichllm/cli.py	311	CODE
LOW	src/whichllm/cli.py	457	CODE
LOW	src/whichllm/cli.py	1239	CODE
LOW	src/whichllm/hardware/gpu_db.py	179	CODE
LOW	src/whichllm/hardware/nvidia.py	130	CODE
LOW	src/whichllm/hardware/cpu.py	14	CODE
LOW	src/whichllm/hardware/cpu.py	100	CODE
LOW	src/whichllm/hardware/cpu.py	134	CODE
LOW	src/whichllm/hardware/cpu.py	171	CODE
LOW	src/whichllm/hardware/gpu_simulator.py	208	CODE
LOW	src/whichllm/hardware/gpu_simulator.py	256	CODE
LOW	src/whichllm/hardware/amd.py	209	CODE
LOW	src/whichllm/output/ranking.py	158	CODE
LOW	src/whichllm/models/benchmark_lineage.py	26	CODE
LOW	src/whichllm/models/benchmark_lookup.py	120	CODE
LOW	src/whichllm/models/grouper.py	62	CODE
LOW	src/whichllm/engine/ranker.py	61	CODE
LOW	src/whichllm/engine/ranker.py	372	CODE
LOW	src/whichllm/engine/ranker.py	474	CODE
LOW	src/whichllm/engine/ranker.py	641	CODE
LOW	src/whichllm/engine/performance.py	140	CODE

Verbosity Indicators4 hits · 9 pts

Severity	File	Line	Snippet	Context
LOW⚡	src/whichllm/cli.py	612	# Step 1: Detect hardware	COMMENT
LOW⚡	src/whichllm/cli.py	621	# Step 2: Fetch models	COMMENT
LOW⚡	src/whichllm/cli.py	641	# Step 3: Fetch benchmark scores	COMMENT
LOW⚡	src/whichllm/cli.py	653	# Step 4: Group and rank	COMMENT

Structural Annotation Overuse4 hits · 9 pts

Severity	File	Line	Snippet	Context
LOW⚡	src/whichllm/cli.py	612	# Step 1: Detect hardware	COMMENT
LOW⚡	src/whichllm/cli.py	621	# Step 2: Fetch models	COMMENT
LOW⚡	src/whichllm/cli.py	641	# Step 3: Fetch benchmark scores	COMMENT
LOW⚡	src/whichllm/cli.py	653	# Step 4: Group and rank	COMMENT

Over-Commented Block6 hits · 6 pts

Severity	File	Line	Snippet	Context
LOW	src/whichllm/models/benchmark_sources/aa_index.py	101		COMMENT
LOW	src/whichllm/data/gpu.py	261		COMMENT
LOW	src/whichllm/data/lineage.py	1	"""Model lineage / generation half-order used to bonus or penalize family versions."""	COMMENT
LOW	src/whichllm/engine/ranker.py	861	)	COMMENT
LOW	src/whichllm/engine/ranker.py	881	# (``results[:-5]``) and silently return a truncated, unrequested subset,	COMMENT
LOW	src/whichllm/engine/performance.py	261		COMMENT

Docstring Block Structure1 hit · 5 pts

Severity	File	Line	Snippet	Context
HIGH	src/whichllm/hardware/gpu_simulator.py	257	Create a synthetic GPUInfo from a GPU name. Looks up specs from the dbgpu database (2000+ GPUs). Args:	STRING

Redundant / Tautological Comments2 hits · 3 pts

Severity	File	Line	Snippet	Context
LOW	src/whichllm/cli.py	720	# Display results	COMMENT
LOW	src/whichllm/engine/ranker.py	838	# Check if this is better than existing	COMMENT

AI Slop Vocabulary1 hit · 3 pts

Severity	File	Line	Snippet	Context
MEDIUM	src/whichllm/engine/ranker.py	75	# MoE offload is more nuanced: inactive experts and router/runtime	COMMENT

AI Structural Patterns2 hits · 2 pts

Severity	File	Line	Snippet	Context
LOW	src/whichllm/cli.py	457		CODE
LOW	src/whichllm/engine/ranker.py	641		CODE

Analysis Overview

What These Metrics Mean

Score History

Severity Breakdown

Directory Score Breakdown

Pattern Findings