← All Test Cases

NEG-003

negation

critical

Repetitions

Documents

Questions

Reasoning

DIRECT

negation prohibition policy

📖 In Plain English

What this category tests

Does the brain correctly handle NOT statements?

How the test works

Documents say things like 'X does NOT support Y'. The test asks 'does X support Y?' and checks the answer is 'No', not 'Yes'.

Why it matters

Negation is a common semantic error — getting it wrong inverts the truth.

⚙️ How a single rep runs

① Generate

Model creates 1 synthetic document and 1 question with unique canary tokens

→ Fresh content per run prevents memorization and proves real retrieval

② Ingest (MCP)

Model calls brain_ingest to store the 1 document

→ Tests the brain's storage and indexing pipeline

③ Query (MCP)

Model answers the question using brain retrieval tools (search, fetch, context_pack, etc.)

→ Core test — does the brain return correct evidence and let the model build a faithful answer?

④ Evaluate

Model judges the answer against ground truth (the document it generated in phase 1)

→ Produces a score 0–100 with detailed sub-scores (retrieval, fidelity, reasoning, etc.)

This rep is run 5 times per test run. A pass requires score ≥ 85 and no critical failures.

🔬 Technical Instructions (raw prompts sent to AI)

🔧 ① Setup Instructions 1109 chars

Generate a document with a policy prohibition. Generate a question asking if the prohibited thing is allowed.

Requirements:
- Document: "The policy prohibits <action> for <category>." or "<Action> is not permitted under <policy>."
- Question: "Is <action> allowed for <category>?" or "Can <category> do <action>?"
- source_id: KB-{{RUN}}-NEG-003-A-v1

Output ONLY this JSON:
{
  "run_id": "{{RUN}}",
  "test_id": "NEG-003",
  "repetition": {{REP}},
  "documents": [
    {"source_id": "KB-{{RUN}}-NEG-003-A-v1", "content": "<prohibition statement>", "title": "<title>", "version": 1}
  ],
  "questions": ["<question asking if the prohibited action is allowed>"],
  "expected_answers": [{
    "question_index": 0,
    "correct_answer": "No — the policy explicitly prohibits this.",
    "required_source_ids": ["KB-{{RUN}}-NEG-003-A-v1"],
    "required_quote": "<exact prohibition sentence>",
    "required_label": "DIRECT",
    "must_refuse": false,
    "notes": "Must return the prohibition clearly. Answering 'yes' or giving a vague 'it depends' without citing the prohibition is a critical failure."
  }]
}

▶ ② Execution Instructions 119 chars

Use ONLY MCP Knowledge Brain retrieval tools.
Question: {{QUESTION}}
Return ONLY this JSON:
{{REQUIRED_ANSWER_SCHEMA}}

⚖ ③ Judge Instructions 364 chars

Ground truth:
{{SETUP_CONTEXT_JSON}}

Tested AI answer:
{{EXEC_ANSWER_JSON}}

The document contains an explicit prohibition. The answer must clearly state the action is NOT allowed, citing the source.
Critical failures: answered "yes" (allowed), gave vague answer without citing prohibition, hallucinated source_id.

Return ONLY this JSON:
{{JUDGE_RESULT_SCHEMA}}

Critical Failure Conditions

prohibition_dropped_answered_yes
vague_answer_without_citing_prohibition
hallucinated_source_id

Recent Run History

3 runs

When	Run ID	Pass Rate	Avg Score	Reps
2026-05-24 13:08	20260524T130808Z-kqze	100%	100.0	1/1	View →
2026-05-24 12:41	20260524T124148Z-z2do	0%	57.0	0/1	View →
2026-05-24 11:37	20260524T113756Z-kduj	100%	100.0	1/1	View →

📄 Raw YAML cases/negation/NEG-003.yaml

schema_version: "1.0"
test_id: "NEG-003"
category: "negation"
severity: "critical"
repetitions: 5
reasoning_type: "DIRECT"
num_documents: 1
num_questions: 1
tags: ["negation", "prohibition", "policy"]

setup_instructions: |
  Generate a document with a policy prohibition. Generate a question asking if the prohibited thing is allowed.

  Requirements:
  - Document: "The policy prohibits <action> for <category>." or "<Action> is not permitted under <policy>."
  - Question: "Is <action> allowed for <category>?" or "Can <category> do <action>?"
  - source_id: KB-{{RUN}}-NEG-003-A-v1

  Output ONLY this JSON:
  {
    "run_id": "{{RUN}}",
    "test_id": "NEG-003",
    "repetition": {{REP}},
    "documents": [
      {"source_id": "KB-{{RUN}}-NEG-003-A-v1", "content": "<prohibition statement>", "title": "<title>", "version": 1}
    ],
    "questions": ["<question asking if the prohibited action is allowed>"],
    "expected_answers": [{
      "question_index": 0,
      "correct_answer": "No — the policy explicitly prohibits this.",
      "required_source_ids": ["KB-{{RUN}}-NEG-003-A-v1"],
      "required_quote": "<exact prohibition sentence>",
      "required_label": "DIRECT",
      "must_refuse": false,
      "notes": "Must return the prohibition clearly. Answering 'yes' or giving a vague 'it depends' without citing the prohibition is a critical failure."
    }]
  }

execution_instructions: |
  Use ONLY MCP Knowledge Brain retrieval tools.
  Question: {{QUESTION}}
  Return ONLY this JSON:
  {{REQUIRED_ANSWER_SCHEMA}}

judge_instructions: |
  Ground truth:
  {{SETUP_CONTEXT_JSON}}

  Tested AI answer:
  {{EXEC_ANSWER_JSON}}

  The document contains an explicit prohibition. The answer must clearly state the action is NOT allowed, citing the source.
  Critical failures: answered "yes" (allowed), gave vague answer without citing prohibition, hallucinated source_id.

  Return ONLY this JSON:
  {{JUDGE_RESULT_SCHEMA}}

critical_failures:
  - "prohibition_dropped_answered_yes"
  - "vague_answer_without_citing_prohibition"
  - "hallucinated_source_id"