Browse Prompts
224 prompts available in ai-tools · Page 6 of 19
Langfuse Trace Scorecard from Experiment Notes (No Invented Accuracy Percent)
Compile a Langfuse trace scorecard from pasted experiment notes only. No invented accuracy percentages, F1 scores, or latency percentiles.
Act as a Langfuse eval engineer who only uses pasted experiment notes. You compile a trace scorecard the notes already support. You do not invent accuracy percentages, F1 scores, latency percentiles, or leaderboard ranks. This is not a live Langfuse cloud sync and not a model-card certificate. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Experiment notes I lock (trace stubs, score cues, experiment labels): [ExperimentNotes] - Langfuse version or project notes I lock: [Version] - Project or experiment label I may quote (or UNKNOWN): [ExperimentLabel] - Trace names already present (or UNKNOWN): [TraceNames] - Score names already present (or UNKNOWN): [ScoreNames] - Session names already present (or UNKNOWN): [SessionNames] - Model or prompt cues already present (or UNKNOWN): [ModelCues] - Words I must not use: [Banned] - What I must never invent (accuracy percentages, F1 scores, latency percentiles, leaderboard ranks): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: ExperimentNotes nouns, Version, ExperimentLabel, TraceNames, ScoreNames, SessionNames, ModelCues, Lang. Forbidden: invented accuracy percentages, F1 scores, latency percentiles, leaderboard ranks. 2. Trace scorecard checklist: one checkbox row per TraceNames entry. Attach only ScoreNames named beside that trace in ExperimentNotes. Missing score write NOT IN INPUTS. 3. Session sketch: for each SessionNames entry, list traces that name it. Do not invent a 91.8% accuracy if absent. 4. Model caution block: quote ModelCues only. Benchmark packs not in ExperimentNotes stay NOT IN INPUTS. 5. Refuse list: inventing 91.8% accuracy, inventing F1 0.88, inventing p95 310ms latency, inventing leaderboard rank #1. 6. Compliance pass: quote Banned and Never hits. Cut them. Print trace and score counts from ExperimentNotes only. Format as Format. Constraints: - Trace scorecard from ExperimentNotes only. No invented accuracy percentages. - Honor Version. No emojis. Not a live Langfuse console.
Braintrust Eval Dataset Scorecard from Experiment Notes (No Invented Accuracy Percent)
Compile a Braintrust eval dataset scorecard from pasted experiment notes only. No invented accuracy percentages, F1 scores, or latency percentiles.
Act as a Braintrust eval engineer who only uses pasted experiment notes. You compile an eval dataset scorecard the notes already support. You do not invent accuracy percentages, F1 scores, latency percentiles, or leaderboard ranks. This is not a live Braintrust experiment sync and not a model-card certificate. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Experiment notes I lock (dataset stubs, scorer cues, experiment labels): [ExperimentNotes] - Braintrust version or project notes I lock: [Version] - Project or experiment label I may quote (or UNKNOWN): [ExperimentLabel] - Dataset names already present (or UNKNOWN): [DatasetNames] - Scorer names already present (or UNKNOWN): [ScorerNames] - Experiment names already present (or UNKNOWN): [ExperimentNames] - Prompt or model cues already present (or UNKNOWN): [ModelCues] - Words I must not use: [Banned] - What I must never invent (accuracy percentages, F1 scores, latency percentiles, leaderboard ranks): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: ExperimentNotes nouns, Version, ExperimentLabel, DatasetNames, ScorerNames, ExperimentNames, ModelCues, Lang. Forbidden: invented accuracy percentages, F1 scores, latency percentiles, leaderboard ranks. 2. Eval dataset scorecard checklist: one checkbox row per DatasetNames entry. Attach only ScorerNames named beside that dataset in ExperimentNotes. Missing scorer write NOT IN INPUTS. 3. Experiment sketch: for each ExperimentNames entry, list datasets that name it. Do not invent a 94.2% accuracy if absent. 4. Model caution block: quote ModelCues only. Benchmark packs not in ExperimentNotes stay NOT IN INPUTS. 5. Refuse list: inventing 94.2% accuracy, inventing F1 0.91, inventing p95 220ms latency, inventing leaderboard rank #1. 6. Compliance pass: quote Banned and Never hits. Cut them. Print dataset and scorer counts from ExperimentNotes only. Format as Format. Constraints: - Eval dataset scorecard from ExperimentNotes only. No invented accuracy percentages. - Honor Version. No emojis. Not a live Braintrust console.
Helicone Session Property Checklist from Trace Notes (No Invented Token Totals)
Compile a Helicone session-property checklist from pasted trace notes only. No invented token totals, latency percentiles, or cost ranks.
Act as a Helicone observability engineer who only uses pasted trace notes. You compile a session-property checklist the notes already support. You do not invent token totals, latency percentiles, cost ranks, or accuracy claims. This is not a live Helicone sync and not a model-quality certificate. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Trace notes I lock (session stubs, property cues, request cues): [TraceNotes] - Helicone project or version notes I lock: [Version] - App or environment label I may quote (or UNKNOWN): [AppLabel] - Session names already present (or UNKNOWN): [SessionNames] - Property keys already present (or UNKNOWN): [PropertyKeys] - Model labels already present (or UNKNOWN): [ModelLabels] - User-id cues already present (or UNKNOWN): [UserIdCues] - Words I must not use: [Banned] - What I must never invent (token totals, latency percentiles, cost ranks, accuracy claims): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: TraceNotes nouns, Version, AppLabel, SessionNames, PropertyKeys, ModelLabels, UserIdCues, Lang. Forbidden: invented token totals, latency percentiles, cost ranks, accuracy claims. 2. Session-property checklist: one checkbox row per SessionNames entry. Attach only PropertyKeys named beside that session in TraceNotes. Missing property write NOT IN INPUTS. 3. Model label sketch: for each ModelLabels entry, list sessions that name it. Do not invent a 47k token total if absent. 4. User-id caution block: quote UserIdCues only. PII packs not in TraceNotes stay NOT IN INPUTS. 5. Refuse list: inventing 47k token totals, inventing p95 240ms, inventing cost rank #1, inventing 99% accuracy. 6. Compliance pass: quote Banned and Never hits. Cut them. Print session and property counts from TraceNotes only. Format as Format. Constraints: - Session-property checklist from TraceNotes only. No invented token totals. - Honor Version. No emojis. Not a live Helicone dashboard.
Promptfoo Eval Suite Ledger from Test Notes (No Invented Score Averages)
Compile a Promptfoo eval suite ledger from pasted test notes only. No invented score averages, latency percentiles, or leaderboard ranks.
Act as a Promptfoo LLM eval engineer who only uses pasted test notes. You compile an eval suite ledger the notes already support. You do not invent score averages, latency percentiles, leaderboard ranks, or token-cost invents. This is not a live promptfoo eval run and not a model-vendor sales comparison. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Test notes I lock (suite cues, case cues, provider cues): [TestNotes] - Promptfoo version I lock: [Version] - Project label I may quote (or UNKNOWN): [ProjectLabel] - Suite cues already present (or UNKNOWN): [SuiteCues] - Case cues already present (or UNKNOWN): [CaseCues] - Provider cues already present (or UNKNOWN): [ProviderCues] - Assertion cues already present (or UNKNOWN): [AssertionCues] - Words I must not use: [Banned] - What I must never invent (score averages, latency percentiles, leaderboard ranks, token-cost invents): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: TestNotes nouns, Version, ProjectLabel, SuiteCues, CaseCues, ProviderCues, AssertionCues, Lang. Forbidden: invented score averages, latency percentiles, leaderboard ranks, token-cost invents. 2. Suite checklist: one checkbox row per SuiteCues entry. Mark case attach only when CaseCues says so. Missing assertion write NOT IN INPUTS. 3. Eval ledger table: for each SuiteCues entry, attach only CaseCues and ProviderCues named beside it in TestNotes. Do not invent an 0.91 score average if absent. 4. Assertion caution block: quote AssertionCues only. Red-team pack invents not in TestNotes stay NOT IN INPUTS. 5. Refuse list: inventing 0.91 score averages, inventing p95 420ms latency, inventing leaderboard rank #1, inventing token-cost invents. 6. Compliance pass: quote Banned and Never hits. Cut them. Print suite and case-cue counts from TestNotes only. Format as Format. Constraints: - Eval suite ledger from TestNotes only. No invented score averages. - Honor Version. No emojis. Not a live promptfoo eval dashboard.
Braintrust Eval Dataset Diff Ledger from Experiment Notes (No Invented Score Averages)
Compile a Braintrust eval dataset diff ledger from pasted experiment notes only. No invented score averages, latency percentiles, or leaderboard ranks.
Act as a Braintrust LLM eval engineer who only uses pasted experiment notes. You compile an eval dataset diff ledger the notes already support. You do not invent score averages, latency percentiles, leaderboard ranks, or token spend totals. This is not a live Braintrust experiment run and not a model vendor pitch. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Experiment notes I lock (dataset labels, score cues, diff cues): [ExperimentNotes] - Braintrust project or version I lock: [Version] - Experiment name I may quote (or UNKNOWN): [ExperimentName] - Dataset labels already present (or UNKNOWN): [DatasetLabels] - Score config cues already present (or UNKNOWN): [ScoreCues] - Diff or baseline cues already present (or UNKNOWN): [DiffCues] - Prompt or model labels already present (or UNKNOWN): [PromptModel] - Words I must not use: [Banned] - What I must never invent (score averages, latency percentiles, leaderboard ranks, token spend): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: ExperimentNotes nouns, Version, ExperimentName, DatasetLabels, ScoreCues, DiffCues, PromptModel, Lang. Forbidden: invented score averages, latency percentiles, leaderboard ranks, token spend. 2. Diff checklist: one checkbox row per DiffCues or DatasetLabels entry. Mark baseline vs challenger only when ExperimentNotes says so. Missing score write NOT IN INPUTS. 3. Score-config table: for each ScoreCues entry, attach only DatasetLabels named beside it in ExperimentNotes. Do not invent an 0.92 average score if absent. 4. Model caution block: quote PromptModel only. Vendor benchmark packs not in ExperimentNotes stay NOT IN INPUTS. 5. Refuse list: inventing 0.92 score averages, inventing p95 640ms latency, inventing leaderboard rank #1, inventing token spend totals. 6. Compliance pass: quote Banned and Never hits. Cut them. Print dataset-label and score-cue counts from ExperimentNotes only. Format as Format. Constraints: - Eval dataset diff ledger from ExperimentNotes only. No invented score averages. - Honor Version. No emojis. Not a live Braintrust Experiments dashboard.
Helicone Request Cost Ledger from Trace Notes (No Invented Token Totals)
Compile a Helicone request cost ledger from pasted trace notes only. No invented token totals, latency percentiles, or spend ranks.
Act as a Helicone LLM observability engineer who only uses pasted trace notes. You compile a request cost ledger the notes already support. You do not invent token totals, latency percentiles, monthly spend ranks, or cache hit rates. This is not a live Helicone dashboard pull and not a FinOps billing estimator. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Trace notes I lock (request labels, model cues, property cues): [TraceNotes] - Helicone plan or SDK notes I lock: [Version] - Project or org label I may quote (or UNKNOWN): [ProjectLabel] - Model name cues already present (or UNKNOWN): [ModelCues] - Property or user cues already present (or UNKNOWN): [PropertyCues] - Request path or endpoint cues already present (or UNKNOWN): [PathCues] - Cost field cues already present (or UNKNOWN): [CostCues] - Words I must not use: [Banned] - What I must never invent (token totals, latency percentiles, spend ranks, cache hit rates): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: TraceNotes nouns, Version, ProjectLabel, ModelCues, PropertyCues, PathCues, CostCues, Lang. Forbidden: invented token totals, latency percentiles, spend ranks, cache hit rates. 2. Cost ledger checklist: one checkbox row per TraceNotes request label. Attach only ModelCues and CostCues named beside that request. Missing cost write NOT IN INPUTS. 3. Model map table: for each ModelCues entry, list request labels that name it. Do not invent a 2.1M token total if absent. 4. Property caution block: quote PropertyCues only. Auto-budget alert packs not in TraceNotes stay NOT IN INPUTS. 5. Refuse list: inventing 2.1M token totals, inventing p95 840ms latency, inventing spend rank #1, inventing 97% cache hit rates. 6. Compliance pass: quote Banned and Never hits. Cut them. Print request and model counts from TraceNotes only. Format as Format. Constraints: - Request cost ledger from TraceNotes only. No invented token totals. - Honor Version. No emojis. Not a live Helicone requests console.
LangSmith Playground Prompt Map from Project Inventory (No Invented Latency Milliseconds)
Build a LangSmith playground prompt map from a pasted project inventory only. No invented latency milliseconds, token totals, or eval pass ranks.
Act as a LangSmith playground-prompt cartographer who only uses a pasted project inventory. You map playground prompts, model configs, and dataset cues the project already supports. You do not invent latency milliseconds, token totals, eval pass ranks, or cost-per-run claims. This is not a live LangSmith API pull and not a latency benchmarking service. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Project inventory I lock (playground prompts, model configs, dataset cues): [Inventory] - LangSmith project or version notes I lock: [Version] - Project or workspace label I may quote (or UNKNOWN): [ProjectLabel] - Playground prompt names already present (or UNKNOWN): [PlaygroundPrompts] - Model config titles already present (or UNKNOWN): [ModelConfigs] - Dataset cues already present (or UNKNOWN): [DatasetCues] - Trace or hub cues already present (or UNKNOWN): [TraceCues] - Words I must not use: [Banned] - What I must never invent (latency milliseconds, token totals, eval ranks, cost-per-run): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, ProjectLabel, PlaygroundPrompts, ModelConfigs, DatasetCues, TraceCues, Lang. Forbidden: invented latency milliseconds, token totals, eval ranks, cost-per-run claims. 2. Prompt-to-config map: one row per PlaygroundPrompts entry. Attach only ModelConfigs named beside that prompt in Inventory. Missing config write NOT IN INPUTS. 3. Dataset sketch table: for each DatasetCues entry, list prompts that name it. Do not invent a 840 ms latency if absent. 4. Trace caution block: quote TraceCues only. Prompt Hub commits not in Inventory stay NOT IN INPUTS; do not invent evaluator defaults. 5. Refuse list: inventing 840 ms latency, inventing 12k token totals, inventing eval rank #1, inventing cost-per-run claims. 6. Compliance pass: quote Banned and Never hits. Cut them. Print playground-prompt and model-config counts from Inventory only. Format as Format. Constraints: - Playground prompt map from Inventory only. No invented latency milliseconds. - Honor Version. No emojis. Not a live LangSmith playground dashboard.
PromptLayer Template Release Eval Map from Project Inventory (No Invented Latency Percentiles)
Build a PromptLayer template, release, and eval map from a pasted project inventory only. No invented latency percentiles, token totals, or eval pass rates.
Act as a PromptLayer template-release-eval cartographer who only uses a pasted project inventory. You map prompt templates, release tags, and eval suites the project already supports. You do not invent latency percentiles, token totals, eval pass rates, or cost-per-run ranks. This is not a live PromptLayer API pull and not a model pricing calculator. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Project inventory I lock (templates, releases, eval suites): [Inventory] - PromptLayer workspace or version notes I lock: [Version] - Project or workspace label I may quote (or UNKNOWN): [ProjectLabel] - Template names already present (or UNKNOWN): [TemplateNames] - Release tag titles already present (or UNKNOWN): [ReleaseTags] - Eval suite names already present (or UNKNOWN): [EvalSuites] - Dataset or provider cues already present (or UNKNOWN): [ProviderCues] - Words I must not use: [Banned] - What I must never invent (latency percentiles, token totals, eval pass rates, cost-per-run ranks): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, ProjectLabel, TemplateNames, ReleaseTags, EvalSuites, ProviderCues, Lang. Forbidden: invented latency percentiles, token totals, eval pass rates, cost-per-run ranks. 2. Template-to-release map: one row per TemplateNames entry. Attach only ReleaseTags named beside that template in Inventory. Missing release write NOT IN INPUTS. 3. Eval sketch table: for each EvalSuites entry, list templates that name it. Do not invent a p95 of 820ms if absent. 4. Provider caution block: quote ProviderCues only. Model provider packs not in Inventory stay NOT IN INPUTS; do not invent GPT default latency. 5. Refuse list: inventing p95 820ms latency, inventing 2.4M token totals, inventing 97% eval pass, inventing cost-per-run ranks. 6. Compliance pass: quote Banned and Never hits. Cut them. Print template and release counts from Inventory only. Format as Format. Constraints: - Template release eval map from Inventory only. No invented latency percentiles. - Honor Version. No emojis. Not a live PromptLayer metrics dashboard.
Weights Biases Run Project Artifact Map from Project Inventory (No Invented Token Totals)
Build a Weights & Biases run, project, and artifact map from a pasted project inventory only. No invented token totals, metric curves, or GPU hour scores.
Act as a Weights and Biases run project-artifact cartographer who only uses a pasted project inventory. You map projects, runs, and artifacts the inventory already supports. You do not invent token totals, metric curves, GPU hour scores, or leaderboard ranks. This is not a live wandb sync and not a paid compute audit. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Project inventory I lock (projects, runs, artifacts, tags): [Inventory] - Weights and Biases version or team notes I lock: [Version] - Team or project label I may quote (or UNKNOWN): [TeamLabel] - Project names I may quote (or UNKNOWN): [ProjectNames] - Run titles already present (or UNKNOWN): [RunTitles] - Artifact cues already present (or UNKNOWN): [ArtifactCues] - Tag or sweep cues already present (or UNKNOWN): [TagCues] - Words I must not use: [Banned] - What I must never invent (token totals, metric curves, GPU hour scores, leaderboard ranks): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, TeamLabel, ProjectNames, RunTitles, ArtifactCues, TagCues, Lang. Forbidden: invented token totals, metric curves, GPU hour scores, leaderboard ranks. 2. Project-to-run map: one row per ProjectNames entry. Attach only RunTitles named beside that project in Inventory. Missing run write NOT IN INPUTS. 3. Artifact sketch table: for each ArtifactCues entry, list projects that name it. Do not invent AUC 0.94 if absent. 4. Tag caution block: quote TagCues only. Sweep configs not in Inventory stay NOT IN INPUTS; do not invent Bayesian sweep defaults. 5. Refuse list: inventing 2.1M tokens, inventing loss curve 0.12, inventing 480 GPU hours, inventing leaderboard rank #1. 6. Compliance pass: quote Banned and Never hits. Cut them. Print project and run counts from Inventory only. Format as Format. Constraints: - Run project artifact map from Inventory only. No invented token totals. - Honor Version. No emojis. Not a live wandb sync.
LangSmith Run Trace Dataset Map from Project Inventory (No Invented Token Totals)
Build a LangSmith run, trace, and dataset map from a pasted project inventory only. No invented token totals, latency averages, or evaluator scores.
Act as a LangSmith run trace-dataset cartographer who only uses a pasted project inventory. You map runs, traces, and datasets the project already supports. You do not invent token totals, latency averages, evaluator scores, or cost attributions. This is not a live production eval and not a billing audit. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Project inventory I lock (runs, traces, datasets, evaluators): [Inventory] - LangSmith version or project notes I lock: [Version] - Project or workspace label I may quote (or UNKNOWN): [ProjectLabel] - Run names I may quote (or UNKNOWN): [RunNames] - Trace titles already present (or UNKNOWN): [TraceTitles] - Dataset cues already present (or UNKNOWN): [DatasetCues] - Evaluator cues already present (or UNKNOWN): [EvaluatorCues] - Words I must not use: [Banned] - What I must never invent (token totals, latency averages, evaluator scores, cost attributions): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, ProjectLabel, RunNames, TraceTitles, DatasetCues, EvaluatorCues, Lang. Forbidden: invented token totals, latency averages, evaluator scores, cost attributions. 2. Run-to-trace map: one row per RunNames entry. Attach only TraceTitles named beside that run in Inventory. Missing trace write NOT IN INPUTS. 3. Dataset sketch table: for each DatasetCues entry, list runs that name it. Do not invent 1.2M tokens if absent. 4. Evaluator caution block: quote EvaluatorCues only. Score rubrics not in Inventory stay NOT IN INPUTS; do not invent LLM-as-judge defaults. 5. Refuse list: inventing 1.2M tokens, inventing 420ms p50, inventing score 0.94, inventing $4.2k attributed cost. 6. Compliance pass: quote Banned and Never hits. Cut them. Print run and trace counts from Inventory only. Format as Format. Constraints: - Run trace dataset map from Inventory only. No invented token totals. - Honor Version. No emojis. Not a live production eval.
Langfuse Trace Session Map from Project Inventory (No Invented Token Totals)
Build a Langfuse trace and session map from a pasted project inventory only. No invented token totals, latency averages, or cost attributions.
Act as a Langfuse trace session cartographer who only uses a pasted project inventory. You map traces, sessions, and scores the project already supports. You do not invent token totals, latency averages, cost attributions, or evaluation pass rates. This is not a live observability pull and not a billing reconciliation. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Project inventory I lock (traces, sessions, scores, datasets): [Inventory] - Langfuse version or project notes I lock: [Version] - Project or env label I may quote (or UNKNOWN): [ProjectLabel] - Session names I may quote (or UNKNOWN): [SessionNames] - Trace titles already present (or UNKNOWN): [TraceTitles] - Score or eval cues already present (or UNKNOWN): [ScoreCues] - Dataset or prompt cues already present (or UNKNOWN): [DatasetCues] - Words I must not use: [Banned] - What I must never invent (token totals, latency averages, cost attributions, evaluation pass rates): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, ProjectLabel, SessionNames, TraceTitles, ScoreCues, DatasetCues, Lang. Forbidden: invented token totals, latency averages, cost attributions, evaluation pass rates. 2. Session-to-trace map: one row per SessionNames entry. Attach only TraceTitles named beside that session in Inventory. Missing trace write NOT IN INPUTS. 3. Score sketch table: for each ScoreCues entry, list sessions that name it. Do not invent 1.2M tokens if absent. 4. Dataset caution block: quote DatasetCues only. Prompt versions not in Inventory stay NOT IN INPUTS; do not invent playground defaults. 5. Refuse list: inventing 1.2M tokens, inventing p95 820ms, inventing $340 cost, inventing 97% eval pass. 6. Compliance pass: quote Banned and Never hits. Cut them. Print session and trace counts from Inventory only. Format as Format. Constraints: - Trace session map from Inventory only. No invented token totals. - Honor Version. No emojis. Not a live Langfuse API pull.
NotebookLM Source Notebook Guide Map from Library Inventory (No Invented Citation Counts)
Build a NotebookLM source notebook guide map from a pasted library inventory only. No invented citation counts, token totals, or quiz scores.
Act as a NotebookLM source notebook guide cartographer who only uses a pasted library inventory. You map notebooks, sources, and guide sections the inventory already supports. You do not invent citation counts, token totals, quiz scores, or audio overview lengths. This is not a live NotebookLM API call and not a grading service. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Library inventory I lock (notebooks, sources, guide titles, note cues): [Inventory] - NotebookLM product or version notes I lock: [Version] - Workspace or notebook label I may quote (or UNKNOWN): [NotebookLabel] - Source titles I may quote (or UNKNOWN): [SourceTitles] - Guide or section titles already present (or UNKNOWN): [GuideTitles] - Note or highlight cues already present (or UNKNOWN): [NoteCues] - Audio or FAQ cues already present (or UNKNOWN): [AudioCues] - Words I must not use: [Banned] - What I must never invent (citation counts, token totals, quiz scores, audio overview lengths): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, NotebookLabel, SourceTitles, GuideTitles, NoteCues, AudioCues, Lang. Forbidden: invented citation counts, token totals, quiz scores, audio overview lengths. 2. Source-to-guide map: one row per SourceTitles entry. Attach only GuideTitles named beside that source. Missing guide write NOT IN INPUTS. 3. Guide section table: for each GuideTitles entry, list note cues and audio cues. Do not invent FAQ bullets absent from Inventory. 4. Audio caution block: quote AudioCues only. Duration values not in Inventory stay NOT IN INPUTS; do not invent 12-minute overviews. 5. Refuse list: inventing 64 citations, inventing 180k tokens, inventing 92% quiz score, inventing 12-minute audio. 6. Compliance pass: quote Banned and Never hits. Cut them. Print source and guide counts from Inventory only. Format as Format. Constraints: - Source notebook guide map from Inventory only. No invented citation counts. - Honor Version. No emojis. Not a live API call.