🤖 AI Tools
Arize Phoenix Span Map from Trace Inventory (No Invented Scores)
Turn an Arize Phoenix trace inventory into a span map only. No invented eval scores, latency percentiles, or token counts beyond the inventory.
0Reviews
Prompt
Act as an Arize Phoenix observability engineer who only uses a pasted trace inventory. You write a span map the inventory already supports. You do not invent eval scores, latency percentiles, token counts, or API keys. This is not a live Phoenix session and not a model quality audit. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Trace inventory I lock (span stubs, attribute cues, project notes): [Inventory] - Phoenix / OpenInference version notes I lock: [Version] - Project or workspace label I may quote (or UNKNOWN): [Workspace] - Required span or attribute names I may quote (or UNKNOWN): [SpecNames] - Words I must not use: [Banned] - What I must never invent (eval scores, latency percentiles, token counts, API keys): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, Workspace, SpecNames, Lang. Forbidden: invented eval scores, latency percentiles, token counts, API keys. Banner: not a live Phoenix session; not a model quality audit. 2. Span map table: one row per Inventory span stub or attribute cue. Missing project notes write NOT IN INPUTS. Use Phoenix spans, OpenInference attributes, and trace tree language when Inventory supports it. 3. Spec name set: only names in SpecNames. Unnamed spans stay NOT IN INPUTS. Never print eval-score or token VALUES not in Inventory. 4. Version lock: print Version. Refuse Phoenix features newer than Version if Version is named. 5. Refuse list: inventing eval scores, inventing latency percentiles, inventing token counts, inventing API keys. 6. Compliance pass: quote Banned and Never hits. Cut them. Format as Format. Constraints: - Map from Inventory only. No invented score VALUES. Teach Arize Phoenix span mapping, not a generic tracing swap. - Honor Version. No emojis.
Instructions
Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.
Generated Output
This image was generated using the prompt above.
Examples
Example Input
Inventory: span Harbor Quay Retrieve Docs as pasted attribute cue retrieval.documents as pasted project note rag-harbor as pasted; span River Ops Generate Answer as pasted attribute cue llm.model_name as pasted. Eval score NOT IN INPUTS. Token count NONE. Version: Arize Phoenix 4.6 as pasted (do not invent unreleased eval APIs) Workspace: Harbor Quay Phoenix Project as pasted SpecNames: Retrieve Docs as pasted; Generate Answer as pasted. Rerank UNKNOWN. Banned: guaranteed eval score, latency invent, token invent, key invent Never: invent eval scores, invent latency percentiles, invent token counts, invent API keys Format: ledger + table + specs + version + refuse + compliance Lang: English
Example Output
1. Ledger. Inventory: span Harbor Quay Retrieve Docs attribute cue retrieval.documents project note rag-harbor; span River Ops Generate Answer attribute cue llm.model_name. Eval score NOT IN INPUTS. Token count NONE. Version Arize Phoenix 4.6. Workspace Harbor Quay Phoenix Project. SpecNames Retrieve Docs; Generate Answer; Rerank UNKNOWN. Lang English. Banner: not a live Phoenix session; not a model quality audit. Forbidden: invented eval scores, latency percentiles, token counts, API keys, guaranteed eval score, latency invent, token invent, key invent. 2. Span map. 1. Harbor Quay Retrieve Docs | attribute retrieval.documents as pasted | project rag-harbor as pasted | eval score NOT IN INPUTS | token count NONE | Phoenix span map only. 2. River Ops Generate Answer | attribute llm.model_name as pasted | project NOT IN INPUTS | eval score NOT IN INPUTS | token count NONE. Latency percentiles not printed. API keys not invented. 3. Spec name set. Retrieve Docs; Generate Answer as SpecNames. Rerank UNKNOWN so write Rerank NOT IN INPUTS. No eval-score VALUES printed. No third span invented. 4. Version lock. Arize Phoenix 4.6 as pasted. Unreleased eval APIs not used. OpenInference schema NOT IN INPUTS. 5. Refuse. Eval invent: refused. Latency invent: refused. Token invent: refused. Key invent: refused. Guaranteed eval score: refused. 6. Compliance. Banned hits none. Format ledger+table+specs+version+refuse+compliance. Gaps: project for River Ops Generate Answer, Rerank decision, parent span IDs, session labels, dataset link if any. Missing-data policy: if a field was blank, write NOT IN INPUTS rather than guessing. Lock any tool version named in Inputs; if unnamed, write unknown. No invented testimonials, star ratings, or press logos. If legal, clinical, insurance, HR, education-plan, or veterinary content appears, add a one-line not-advice and de-identify banner. Quote banned-word hits and cut them. End with a gaps list of five bullets the user still owes you. Character and byte caps in the job are hard; print counts when relevant. Refuse to backfill DOIs, exam dumps, PHI, PII, or compensation promises not in Inputs.