Does Dify answer when the material does not support an answer?
Dify handled the unsupported question by saying it did not know and giving no leave amount or policy. The retrieved passages were unrelated to parental leave and instead covered handbook and field-service material.
1 of 1 test case passed
Every test case in this scenario has a result.
The test case
Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.
Answer generated when the material does not support onePassEvidence
“How much paid parental leave does Fenwake give?”
I don’t know. I don’t have Fenwake’s parental leave policy details. Could you clarify which type of leave you mean (maternity, paternity/partner, adoption, shared parental) and your location? For the exact paid entitlement, please check your written statement of particulars or the Family/Parental Leave section of the employee handbook, or contact HR.
Transcribed from Proof 1 · Screenshot
POST /chat-messages Request body: { "query": "How much paid parental leave does Fenwake give?", "inputs": {}, "user": "r6-tester", "response_mode": "blocking" } Status: 200 Response body: { "event": "message", "task_id": "7a86f0d3-a91f-4dca-8248-8d33ff327b78", "id": "1fd23db1-1ba3-449e-9f8b-c0b9b2d2522c", "message_id": "1fd23db1-1ba3-449e-9f8b-c0b9b2d2522c", "conversation_id": "0aa536a4-19dd-4d66-9486-1d5431997870", "mode": "chat", "answer": "I don\u2019t know. I don\u2019t have Fenwake\u2019s parental leave policy details.\n\nCould you clarify which type of leave you mean (maternity, paternity/partner, adoption, shared parental) and your location? For the exact paid entitlement, please check your written statement of particulars or the Family/Parental Leave section of the employee handbook, or contact HR.", "metadata": { "annotation_reply": null, "retriever_resources": [ { "position": 1, "dataset_id": "084fc635-481f-4970-bf09-299c83eca3c2", "dataset_name": "R6-managed-rag-ss-v1", "document_id": "4c1c336c-11c5-423f-bf1e-71d339564495", "document_name": "A04_employee_handbook.docx", "data_source_type": "upload_file", "segment_id": "c34fb06b-a2cd-4d29-aec6-6b4ae2b12c6f", "retriever_from": "api", "score": 0.49496233, "hit_count": null, "word_count": null, "segment_position": null, "index_node_hash": null, "content": "Employee Handbook\nFenwake Instruments Ltd - all locations\n1. Working here\nThis handbook sets out the terms that apply to everyone employed by Fenwake Instruments Ltd, whether based at the Bristol workshop, at the Avonmouth calibration laboratory, or working from a customer site or vessel. It forms part of your contract. Where a term is set out in your written statement of particulars and also here, the written statement governs.\n2.", "page": null, "doc_metadata": { "region": "group" }, "title": null, "files": null, "summary": null }, { "position": 2, "dataset_id": "084fc635-481f-4970-bf09-299c83eca3c2", "dataset_name": "R6-managed-rag-ss-v1", "document_id": "0eeae560-0e74-44e9-aaee-2cc6b599dee8", "document_name": "A01_field_service_policy_and_charges_uk.pdf", "data_source_type": "upload_file", "segment_id": "f5751516-6078-4011-8674-34dd47583750", "retriever_from": "api", "score": 0.48522162, "hit_count": null, "word_count": null, "segment_position": null, "index_node_hash": null, "content": "Fenwake Instruments Ltd - Field Service Policy and Charges (United Kingdom)\r\nFenwake internal - service operations Page 2\r\n4. Acknowledgement during business hours\r\nBusiness hours are 08:00 to 18:00, Monday to Friday, excluding statutory non-working days in England\r\nand Wales.", "page": null, "doc_metadata": { "region": "uk" }, "title": null, "files": null, "summary": null }, { "position": 3, "dataset_id": "084fc635-481f-4970-bf09-299c83eca3c2", "dataset_name": "R6-managed-rag-ss-v1", "document_id": "0eeae560-0e74-44e9-aaee-2cc6b599dee8", "document_name": "A01_field_service_policy_and_charges_uk.pdf", "data_source_type": "upload_file", "segment_id": "5d1c8dd2-8d4d-44ab-a6d2-dc6ac2ac315a", "retriever_from": "api", "score": 0.47646844, "hit_count": null, "word_count": null, "segment_position": null, "index_node_hash": null, "content": "Fenwake internal - service operations Page 1\r\nFenwake Instruments Ltd\r\nField Service Policy and Charges\r\nUnited Kingdom operations\r\n1. Purpose\r\nThis document states what a customer holding a support contract with Fenwake Instruments Ltd may\r\nexpect when an instrument fails in service, what the service desk in Bristol is obliged to do in response,\r\nand what is charged for work that falls outside cover.", "page": null, "doc_metadata": { "region": "uk" }, "title": null, "files": null, "summary": null }, { "position": 4, "dataset_id": "084fc635-481f-4970-bf09-299c83eca3c2", "dataset_name": "R6-managed-rag-ss-v1", "document_id": "0eeae560-0e74-44e9-aaee-2cc6b599dee8", "document_name": "A01_field_service_policy_and_charges_uk.pdf", "data_source_type": "upload_file", "segment_id": "aca8442b-62c2-4e39-99c8-a72352e26c07", "retriever_from": "api", "score": 0.4744957, "hit_count": null, "word_count": null, "segment_position": null, "index_node_hash": null, "content": "Fenwake Instruments Ltd - Field Service Policy and Charges (United Kingdom)\r\nFenwake internal - service operations Page 6\r\n8. Contacts\r\nRoute When Detail\r\nService desk Business hours service@fenwake-instruments.example\r\nService desk telephone Business hours +44 117 496 0188\r\nOut-of-hours pager All other times +44 117 496 0199\r\n9.", "page": null, "doc_metadata": { "region": "uk" }, "title": null, "files": null, "summary": null } ], "usage": { "prompt_tokens": 401, "prompt_unit_price": "1.25", "prompt_price_unit": "0.000001", "prompt_price": "0.0005013", "completion_tokens": 552, "completion_unit_price": "10", "completion_price_unit": "0.000001", "completion_price": "0.00552", "total_tokens": 953, "total_price": "0.0060213", "currency": "USD", "latency": 8.439466020092368, "time_to_first_token": null, "time_to_generate": null }, "reasoning": {} }, "created_at": 1789934048 } App configuration actually applied (read live from the R6-managed-rag-chatbot app's Orchestrate panel -- no API endpoint exposes this, per reference/api-dify.md, so this is a UI-sourced fact, not assumed -- same app and config as TC94/TC96/TC97, confirmed once): Generation model: gpt-5 (Chat mode). Corroborated by the usage block above: prompt_unit_price 1.25 / completion_unit_price 10 (per 1M tokens) match OpenAI's published gpt-5 pricing. Retrieval config (Orchestrate -> Knowledge -> Retrieval Setting): - Strategy: multi-path retrieval (default) -- retrieved from R6-managed-rag-ss-v1, then reranked - Rerank setting: Weighted Score -- Semantic 1.0 / Keyword 0.0 (fully semantic weighting, no keyword component) - Top K: 4 - Score Threshold: disabled (off; the 0.7 field value shown is not applied) This lets a reader judge the decline as well-configured, not a lucky miss: with a score threshold disabled, all 4 near-miss passages (max score 0.495) were still surfaced to the model rather than silently filtered out -- the decline came from the model correctly reading that none of them address parental leave, not from retrieval returning nothing.
| ✓ | Expected: The answer should say the material does not cover the question.Found: It says it does not know, gives no leave amount or policy, and the retrieved passages stay on unrelated handbook and field-service topics. |
It declines instead of inventing a policy, and the retrieved passages stay on unrelated handbook and field-service content, so the unsupported question is handled correctly.
Configuration and setup
How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.
How this scenario is graded
How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.
Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.
- Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
- Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
- Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.
Where this sits in the benchmark
This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.
| Level | Name | Scope |
|---|---|---|
| Benchmark | Managed RAG → | 15 scenarios · 7 tools |
| Capability | Grounded Generation → | |
| Scenario | The retrieved material doesn't support an answer → | |
| Tool | Dify → |
Global scenario definition → · Global capability definition → · Dify product page →
History of this result
What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.
Act on this result
Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.
You run the same kind of test against your own setup and get the same behaviour.
Agree →Yours behaves differently. Tell us what you got, with a screenshot if you have one.
Disagree →Something here is wrong — a reference value, a transcription, a grade.
Report an issue →Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.
Vendor notice →The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.
Verify the proof files
These files support this result. Open a file to inspect the original evidence.
File fingerprints (SHA-256)
A fingerprint identifies the exact file used for this result.