MCPcopy Create free account

hub / github.com/19PINE-AI/ikp / functions

Functions549 in github.com/19PINE-AI/ikp

⨍ Functions 549◇ Types & classes 37
↓ 36 callersMethodadd
Add a probe. Returns its ID, or None if duplicate.
pipeline/store.py:65
↓ 21 callersFunctionvendorColor
(vendor: string | null | undefined)
website/src/util.ts:57
↓ 14 callersFunctionformatPercent
(v: number, decimals = 1)
website/src/util.ts:1
↓ 14 callersFunctionlog
(msg)
scripts/rerun_researcher_probes.py:54
↓ 14 callersFunctionscaleY
(v: number)
website/src/pages/TierHeatmap.tsx:177
↓ 14 callersFunctionuseFetch
(path: string)
website/src/data.ts:34
↓ 13 callersMethodget_response_text
Convenience: return just the text content of the first choice.
src/api_client.py:144
↓ 8 callersMethodsave
(self)
pipeline/store.py:61
↓ 8 callersFunctionvendor_of
(name: str)
scripts/14_comprehensive_fingerprinting.py:83
↓ 7 callersFunctionescape_latex
(s)
scripts/14_comprehensive_fingerprinting.py:512
↓ 7 callersFunctiong
(n)
scripts/lambda_floor_ablation.py:89
↓ 7 callersFunctionsave
(fig, name)
paper/figures/generate_figures.py:125
↓ 7 callersFunctiontex_escape
Escape special LaTeX characters in a string.
paper/figures/generate_figures.py:116
↓ 6 callersFunctioncolorFor
(cls: string)
website/src/pages/Fingerprint.tsx:40
↓ 6 callersFunctionis_refusal
Detect if the model refused to answer.
src/scorer.py:18
↓ 6 callersFunctionq
(p: number)
website/src/pages/TierHeatmap.tsx:23
↓ 6 callersFunctiontest_coef
(model, idx, target)
scripts/15_densing_law_analysis.py:195
↓ 6 callersFunctionvendor_of
(name)
paper/figures/generate_figures.py:852
↓ 5 callersFunctionfit_ols
(df: pd.DataFrame, X_cols: list[str], y_col: str = "pen_acc")
scripts/15_densing_law_analysis.py:132
↓ 5 callersFunctionformatBigParams
(b: number | null | undefined)
website/src/pages/Calibration.tsx:276
↓ 5 callersMethodget_stats
(self)
src/api_client.py:153
↓ 5 callersFunctionnice_coef_table
(model, names)
scripts/15_densing_law_analysis.py:138
↓ 5 callersMethodprint_status
Print a human-readable status report.
pipeline/store.py:193
↓ 5 callersFunctiontoggleSort
( k: SortKey, current: SortKey, dir: "asc" | "desc", setKey: (k: SortKey) => void, setDir: (d: "asc"
website/src/pages/Models.tsx:179
↓ 4 callersFunction_pair_id
Stable URL-safe id for an (a, b) pair, normalised so that pair_id(a, b) == pair_id(b, a). Frontend uses this id as the route parameter and as
website/scripts/prepare_data.py:706
↓ 4 callersFunctionfit
(items, key="params_B")
scripts/moe_dense_analysis.py:23
↓ 4 callersFunctionformatBigParams
(b: number | null | undefined)
website/src/pages/Home.tsx:385
↓ 4 callersFunctionformatParams
(b: number | null | undefined)
website/src/util.ts:5
↓ 4 callersFunctionload_probes
Load probes from a JSON file.
src/probe_generator.py:278
↓ 4 callersFunctionnorm_inst
(a)
scripts/clean_flag_researchers.py:21
↓ 4 callersFunctionsave
(fig, name)
paper/figures/generate_appendix_figures.py:99
↓ 4 callersFunctionsave_probes
Save probes to a JSON file.
src/probe_generator.py:267
↓ 4 callersFunctionuseModels
()
website/src/data.ts:58
↓ 4 callersFunctionvalidate_probe_format
Validate that a probe has all required fields.
src/probe_generator.py:251
↓ 3 callersFunctionacc
(p)
scripts/moe_dense_analysis.py:22
↓ 3 callersMethodadd_batch
Add multiple probes. Returns count of new (non-duplicate) probes added.
pipeline/store.py:81
↓ 3 callersFunctionadd_const
(X)
scripts/15_densing_law_analysis.py:67
↓ 3 callersFunctionevaluate_model
Evaluate a single model against all probes.
scripts/run_evaluation.py:266
↓ 3 callersFunctionfit_joint
(x_p, x_m, y)
website/scripts/prepare_data.py:1097
↓ 3 callersFunctionfit_ols
(x: np.ndarray, y: np.ndarray)
scripts/16_benchmark_comparison.py:39
↓ 3 callersFunctionfit_simple
(xs, ys)
website/scripts/prepare_data.py:1092
↓ 3 callersFunctionfit_with_time
OLS y ~ log_params + months. Returns slope on params, slope on time, R^2, n.
scripts/16_benchmark_comparison.py:51
↓ 3 callersFunctionfmt
(v: number | undefined | null, digits: number)
website/src/pages/Fingerprint.tsx:262
↓ 3 callersFunctionformat_params
Format parameter count for display.
scripts/ikp_estimate.py:87
↓ 3 callersFunctionhead
(k: typeof sortKey, label: string)
website/src/pages/Recognition.tsx:208
↓ 3 callersFunctionjudge
Return one of CORRECT, WRONG, REFUSAL.
cli/judge.py:17
↓ 3 callersFunctionload_all_probes
Load probes from all tiers.
src/probe_runner.py:230
↓ 3 callersFunctionnormalize_probe
Normalize a probe to a consistent schema.
scripts/assemble_final_dataset.py:41
↓ 3 callersFunctionparamLabel
(v: number)
website/src/pages/Calibration.tsx:575
↓ 3 callersFunctionrun_model_probes
Run all probes against a single model. Resumable. Returns a summary dict with per-tier accuracies.
src/probe_runner.py:112
↓ 2 callersFunction_call_judge
(prompt, out, judge_model, api_key, http, valid_verdicts=("CORRECT_STRONG", "CORRECT_WEAK", "W
src/evidence_judge.py:155
↓ 2 callersFunction_parse_json_response
Parse JSON array from LLM response, handling common formatting issues.
scripts/01b_generate_t6_t7.py:186
↓ 2 callersFunction_print_header
(question: str, gold: str | None)
cli/research.py:16
↓ 2 callersFunction_run_against_models
(question: str, models: list)
cli/research.py:40
↓ 2 callersFunction_save_raw_results
(output_file: Path, model_name: str, model_id: str, probe_results: list)
src/probe_runner.py:170
↓ 2 callersFunction_seed_pair
(a: str, b: str, source: str, extra: dict | None = None)
website/scripts/prepare_data.py:827
↓ 2 callersFunction_strip_tool_handlers
Remove tool-call / invoke wrappers some models emit when they treat the factual prompt as a tool-use task. Strips patterns like '<minimax:tool
website/scripts/prepare_data.py:351
↓ 2 callersFunction_wrapped_rows
How many visual rows the string occupies when the terminal wraps at term_w.
cli/progress.py:41
↓ 2 callersFunctionacc
(ts, lam, floor=True)
paper/figures/generate_lambda_figure.py:34
↓ 2 callersFunctionacc
(ts, lam, floor)
scripts/lambda_floor_ablation.py:36
↓ 2 callersFunctionacc_at
(tier_stats, lam)
scripts/lambda_sensitivity.py:42
↓ 2 callersFunctionassemble_one
Recompute final-assembled accuracy at a custom lambda.
scripts/calibration_refit_v2.py:47
↓ 2 callersFunctionbuild_generation_prompt
Build a structured prompt to generate probe candidates for a given tier.
src/probe_generator.py:117
↓ 2 callersMethodchat_completion
Send a chat completion request. Returns the full API response dict.
src/api_client.py:77
↓ 2 callersFunctioncreate_openrouter_judge
Create a judge function using an OpenRouter model. Uses direct httpx calls (no rate limiter) for parallel execution.
src/scorer.py:115
↓ 2 callersFunctiondwords
(s)
scripts/clean_flag_wikidata.py:64
↓ 2 callersFunctionesc
(s)
scripts/full_results_tables.py:19
↓ 2 callersFunctionestimate_log_params_aggregate
Estimate log(N) from aggregate accuracy using log-linear fit.
src/calibration.py:116
↓ 2 callersFunctionestimate_log_params_per_tier
Estimate log(N) from per-tier accuracies using logistic inversion. Uses weighted median where weight = information content (highest near 50%
src/calibration.py:121
↓ 2 callersFunctionestimate_params
Estimate parameter count from tier accuracies using fitted curves. Returns geometric mean of per-tier estimates.
scripts/analyze_results.py:77
↓ 2 callersFunctionestimate_params
Estimate parameter count in billions from penalized accuracy.
scripts/ikp_estimate.py:79
↓ 2 callersFunctionf
(v)
scripts/frontier_table.py:56
↓ 2 callersFunctionfetch_wikipedia_random_articles
Fetch random Wikipedia articles to discover obscure facts.
scripts/11_corpus_grounded_t5t7.py:210
↓ 2 callersFunctionfinalize
(i: int)
cli/progress.py:130
↓ 2 callersFunctionfit_aggregate_log_linear
Fit log(N) = alpha * accuracy + beta using OLS. Returns: (alpha, beta, r_squared, rmse)
src/calibration.py:31
↓ 2 callersFunctionfit_moe
Fit log-linear accuracy ~ log10(params) for MoE models on either total or active.
website/scripts/prepare_data.py:120
↓ 2 callersFunctionfit_tier_logistic
Fit a logistic sigmoid to one tier's accuracy vs log(params). T_i(log_N) = L / (1 + exp(-k * (log_N - m))) Returns: dict with fitted paramet
src/calibration.py:61
↓ 2 callersFunctionfmtTime
(v: number)
website/src/pages/Densing.tsx:401
↓ 2 callersFunctionfmt_size
(b)
scripts/lambda_sensitivity.py:52
↓ 2 callersFunctionfold
(pred, act)
scripts/moe_dense_analysis.py:55
↓ 2 callersFunctionfp
(b)
scripts/full_results_tables.py:20
↓ 2 callersFunctionfs
(b)
scripts/moe_dense_analysis.py:70
↓ 2 callersFunctionfsize
(b)
scripts/frontier_table.py:51
↓ 2 callersFunctiongenerate_all
(n_per_type: int = 200)
pipeline/generate_wikidata.py:140
↓ 2 callersFunctiongenerate_tier_probes
Generate probe candidates for a single tier.
src/probe_generator.py:173
↓ 2 callersFunctionget
(params)
scripts/clean_flag_wikidata.py:25
↓ 2 callersFunctionjudge_evidence
Judge a researcher response against an enriched gold bundle. Returns dict with keys: verdict, judge_prompt, judge_raw_output, error. Verdict
src/evidence_judge.py:95
↓ 2 callersFunctionjudge_fn
(prompt: str)
src/scorer.py:97
↓ 2 callersFunctionload_per_model_results
(model_name)
website/scripts/prepare_data.py:60
↓ 2 callersFunctionload_probes
()
cli/probes.py:10
↓ 2 callersFunctionlogN
(p, key="params_B")
scripts/moe_dense_analysis.py:21
↓ 2 callersFunctionlogistic
Logistic sigmoid: L / (1 + exp(-k*(x - m)))
src/calibration.py:26
↓ 2 callersFunctionloo_pred
(target, pool, key="params_B")
scripts/moe_dense_analysis.py:50
↓ 2 callersFunctionmain
()
cli/main.py:14
↓ 2 callersFunctionnormalizeClass
(c: string | undefined)
website/src/pages/Fingerprint.tsx:35
↓ 2 callersFunctionpairIdFor
(p: FingerprintPair)
website/src/pages/Fingerprint.tsx:323
↓ 2 callersFunctionpick
(m: ModelSummary, k: SortKey)
website/src/pages/Models.tsx:173
↓ 2 callersFunctionpretty
(name)
scripts/frontier_table.py:32
↓ 2 callersFunctionprobe_id
Deterministic ID from question text.
pipeline/store.py:39
↓ 2 callersFunctionquery_model
Query a single model spec (see presets.LANDMARKS for shape).
cli/query.py:101
next →1–100 of 549, ranked by callers