806 lines
50 KiB
JSON
806 lines
50 KiB
JSON
{
|
||
"calibration_version": "0.3-draft",
|
||
"status": "draft_owner_review_needed",
|
||
"notes": "These inputs calibrate selector behavior only. They are not model regression cases and do not replace owner-supplied raw QPI case materials.",
|
||
"inputs": [
|
||
{
|
||
"case_id": "selector_calibration_fact_lookup_001",
|
||
"category": "fact_lookup",
|
||
"input": "查一下 QPI 这个缩写是什么意思，只要事实。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Hard no-call because the user asks for lookup only."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_fact_lookup_002",
|
||
"category": "fact_lookup",
|
||
"input": "这篇文章的发布日期是哪天？不要展开。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Date lookup does not need QPI or IA."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_rewrite_001",
|
||
"category": "rewrite_polish_translation",
|
||
"input": "请直接润色这段话，不要深入分析。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Pure polish request."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_translation_001",
|
||
"category": "rewrite_polish_translation",
|
||
"input": "翻译这句话：The model is provisional.",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Translation request."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_direct_execution_001",
|
||
"category": "direct_execution",
|
||
"input": "把这个 JSON 缩进格式化一下，马上执行，不用解释。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Direct execution request."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_only_001",
|
||
"category": "qpi_only",
|
||
"input": "我们现在的问题到底是缺数据、缺路径，还是缺共识？",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Explicit QPI problem-definition gate."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_only_002",
|
||
"category": "qpi_only",
|
||
"input": "这个需求反复卡住，我不确定该怎么定义这个问题。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Problem definition is incomplete."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_insufficient_context_001",
|
||
"category": "mixed_qpi_cases",
|
||
"input": "如何提高流量？",
|
||
"expected_selector_behavior": "select_qpi_low_confidence",
|
||
"expected_notes": "Selector may route to QPI, but QPI output must be provisional because subject and scenario are missing."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_intra_frame_mixed_001",
|
||
"category": "mixed_qpi_cases",
|
||
"input": "我是集团营销负责人，既缺流量增长方法，也担心销售、库存、客服和激励机制跟不上。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Same subject and scenario contain path/resource and governance scarcity."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_governance_load_001",
|
||
"category": "mixed_qpi_cases",
|
||
"input": "这个流程改造表面只是工具不熟，但实际牵涉真实调用、上下文裁剪、审计边界和后续复用。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Governance load raises Issue weight even if the owner is one person."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_ia_after_qpi_001",
|
||
"category": "ia_after_qpi",
|
||
"input": "QPI 已判断这是中重型课题，请用思想考古分析它的底层假设和核心机制。",
|
||
"expected_selector_behavior": "select_intellectual_archaeology",
|
||
"expected_notes": "QPI-before-IA gate is satisfied."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_ia_after_qpi_002",
|
||
"category": "ia_after_qpi",
|
||
"input": "QPI 已判断这是 P/I mixed，请下潜到 purpose 或 core_mechanism，说明不用到哲学基岩的理由。",
|
||
"expected_selector_behavior": "select_intellectual_archaeology",
|
||
"expected_notes": "Depth should be gated, not default full seven-layer analysis."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_ia_model_extraction_001",
|
||
"category": "ia_explicit_model_extraction",
|
||
"input": "请从这篇文章抽取一个可复用模型，并检查它的底层假设是否稳定。",
|
||
"expected_selector_behavior": "select_intellectual_archaeology",
|
||
"expected_notes": "Explicit model extraction and heavy-depth signal."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_ia_model_extraction_002",
|
||
"category": "ia_explicit_model_extraction",
|
||
"input": "我想把这套专家直觉沉淀成模型，需要识别隐性知识和验证点。",
|
||
"expected_selector_behavior": "select_intellectual_archaeology",
|
||
"expected_notes": "Expert intuition and tacit knowledge are IA triggers."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_false_positive_deep_word_001",
|
||
"category": "false_positive_traps",
|
||
"input": "把这句话里的“底层逻辑”换成更通俗的说法，不要深入分析。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Deep word appears, but hard no-call rewrite intent wins."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_false_positive_model_word_001",
|
||
"category": "false_positive_traps",
|
||
"input": "翻译 model 这个词，不要展开。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "The word model is not a model-extraction request."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_false_positive_philosophy_001",
|
||
"category": "false_positive_traps",
|
||
"input": "查一下哲学基岩这个词出现在哪篇文章，只查位置。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Lookup request wins over IA keyword."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_override_001",
|
||
"category": "direct_execution_with_analysis_override",
|
||
"input": "这看起来只是润色，但不要只执行，请分析背后的问题结构。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Explicit analysis override allows QPI despite a rewrite-like surface."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_pipeline_001",
|
||
"category": "pipeline",
|
||
"input": "问题定义还没完成，先不要思想考古，先判断这是 Q、P 还是 I。",
|
||
"expected_selector_behavior": "select_qpi_reject_ia",
|
||
"expected_notes": "QPI-before-IA gate."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_pipeline_002",
|
||
"category": "pipeline",
|
||
"input": "QPI 已判断不是事实查询，而是系统性课题；下一步检查它是否值得沉淀成可复用模型。",
|
||
"expected_selector_behavior": "select_intellectual_archaeology",
|
||
"expected_notes": "QPI has completed the front gate."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_flow_entry_point_001",
|
||
"category": "qpi_owner_reviewed_flow_redesign",
|
||
"input": "我们本来想自动化大纲审阅，但发现真正卡住的是原始材料压缩、前置上下文、主笔框定和 owner 确认，review 只是很晚才暴露问题。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed flow-redesign case 001: wrong automation entry point reveals context engineering and stage authority scarcity."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_no_simulation_001",
|
||
"category": "qpi_owner_reviewed_flow_redesign",
|
||
"input": "这个多代理评审包看起来有完整报告，但其实只是主会话模拟，没有真实独立 agent 调用记录；我需要判断它能不能作为正式输出。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed flow-redesign case 002: no-simulation and formal-output legitimacy should route to QPI issue framing."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_local_truth_global_structure_001",
|
||
"category": "qpi_owner_reviewed_flow_redesign",
|
||
"input": "分块提炼都局部正确，但整体主轴、主题结构和模型演化线被误读了；我怀疑局部真实冒充了全局结构。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed flow-redesign case 003: local truth impersonating global structure is a QPI misframing risk."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_stop_gate_001",
|
||
"category": "qpi_owner_reviewed_flow_redesign",
|
||
"input": "素材提炼完成后，不应该自动创建写作项目或自动进入模型挖掘；下一步必须先经过 owner 授权和 source-of-truth 确认。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed flow-redesign case 004: stop gates and downstream transition authority are QPI issue signals."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_intra_frame_mixed_flow_001",
|
||
"category": "qpi_owner_reviewed_flow_redesign",
|
||
"input": "这只是我个人的流程改造，但它涉及真实调用、上下文裁剪、审计边界、token 成本、后续复用和未来 owner 的维护责任。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed flow-redesign case 005: one-owner workflow can be intra_frame_mixed when governance load is sustained."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_complexity_placement_001",
|
||
"category": "qpi_owner_reviewed_flow_redesign",
|
||
"input": "我不想简单判断这个流程是不是过度工程，而是要判断这些复杂度分别买到了真实性、保真性、可追溯性、可停止性和 ROI 吗。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed flow-redesign case 006: complexity placement gate is QPI issue calibration, not a third model."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_psych_mechanism_ambiguity_001",
|
||
"category": "qpi_owner_reviewed_disappointment_isolation",
|
||
"input": "同一个“失望性情感隔离”标签，可能是被动麻木，也可能是主动认知抽离；我需要先区分机制再判断这是什么问题。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed disappointment case 001: mechanism ambiguity and multi-perspective routing should select QPI."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_label_as_identity_001",
|
||
"category": "qpi_owner_reviewed_disappointment_isolation",
|
||
"input": "我不是只想查这个心理标签是什么意思，它已经影响我怎么理解自己、怎么设关系边界和要不要寻求帮助。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed disappointment case 002: a label used for identity and relationship governance is not pure Q lookup."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_single_factor_totalization_001",
|
||
"category": "qpi_owner_reviewed_disappointment_isolation",
|
||
"input": "我不想把情感隔离全部归因于童年或某一个人，我需要判断哪些是触发、维持条件、高权重要素和当前行动杠杆。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed disappointment case 003: single-factor totalization is a QPI misframing risk."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_one_time_fix_trap_001",
|
||
"category": "qpi_owner_reviewed_disappointment_isolation",
|
||
"input": "这个心理关系问题不能靠三步修复或一次干预解决，我需要判断它是不是只能阶段性重构和长期治理。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed disappointment case 004: one-time solution framing should route to QPI issue diagnosis."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_external_authority_boundary_001",
|
||
"category": "qpi_owner_reviewed_disappointment_isolation",
|
||
"input": "我想判断外部帮助是在替我给答案，还是帮我建立自我治理和校准环境，这涉及权威边界和依赖风险。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed disappointment case 005: external authority boundary is a governance-load signal."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_time_scale_scope_001",
|
||
"category": "qpi_owner_reviewed_disappointment_isolation",
|
||
"input": "同一个关系现象，在概念层像 Q，在一次行动上像 P，在长期生命模式上又像 I；我需要先确定时间尺度和分类范围。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed disappointment case 006: time scale and classification scope determine QPI routing."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_mismatch_diagnostics_001",
|
||
"category": "qpi_owner_reviewed_disappointment_isolation",
|
||
"input": "我需要检查这个心理关系材料有没有被误路由：把 I 当 Q、把 I 当 P、把 P 过度升维成 I，或者把保护性边界病理化。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed disappointment case 007: QPI mismatch diagnostics should select QPI."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_one_person_issue_001",
|
||
"category": "qpi_owner_reviewed_disappointment_isolation",
|
||
"input": "只有我一个人在做自我治理，但我要长期校准隔离什么时候是保护，什么时候是过度泛化防御，这是不是仍然可以是 Issue？",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed disappointment case 008: one-person self-governance can still be Issue."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_org_hard_resource_documentation_001",
|
||
"category": "qpi_owner_reviewed_year_end_development_office",
|
||
"input": "组织面对硬性资源缺口，没有真实补资源，而是想通过重命名、材料包装和合规口径让指标看起来达标。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed development-office case 001: hard resource gap disguised as documentation problem should route to QPI as governance/order scarcity with path/resource trigger."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_org_credential_continuity_001",
|
||
"category": "qpi_owner_reviewed_year_end_development_office",
|
||
"input": "为了修复人员资质指标，组织准备替换或重分类一批岗位，但这些人承担运营记忆、角色连续性和信任关系。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed development-office case 002: credential metric repair damaging operational continuity is QPI issue / violent reduction."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_org_incentive_backlash_001",
|
||
"category": "qpi_owner_reviewed_year_end_development_office",
|
||
"input": "管理层把绩效下滑解释成员工不努力，但一线实际是在负激励和成本转嫁下理性减少投入。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed development-office case 003: performance decline with incentive backlash should not be reduced to effort problem."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_org_role_authority_mismatch_001",
|
||
"category": "qpi_owner_reviewed_year_end_development_office",
|
||
"input": "一个战略支持职能被要求拿出战略增长和外部合作，却没有资源、授权、信任和可执行空间，最后只能做指标包装。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed development-office case 004: strategic role without strategic operating conditions is QPI role-authority mismatch."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_org_pattern_level_issue_001",
|
||
"category": "qpi_owner_reviewed_year_end_development_office",
|
||
"input": "同一组织反复要求高目标低投入：要合作但不给交换，要产出但不给激励，要容量但不扩资源，下游只能吸收矛盾。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed development-office case 005: pattern-level Issue from shared low-cost ambition / target-resource inconsistency mechanism."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_org_untrusted_data_001",
|
||
"category": "qpi_owner_reviewed_year_end_development_office",
|
||
"input": "材料和台账都很完整，但它们对应的真实运营并不稳定，也经不起审计核查；这不是缺数据，而是数据不可信。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed development-office case 006: data completeness versus truthfulness gap should route to QPI evidence authenticity issue."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_org_workaround_translation_001",
|
||
"category": "qpi_owner_reviewed_year_end_development_office",
|
||
"input": "中层执行者知道指令和现实冲突，但仍把不可能的要求翻译成可运行的变通动作，让错误治理逻辑继续运转。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed development-office case 007: role-boundary governance and workaround normalization."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_org_low_cost_indicator_survival_001",
|
||
"category": "qpi_owner_reviewed_year_end_development_office",
|
||
"input": "外部指标压力只是触发器，真正的问题是组织内部低投入、高压和形式主义治理，把指标生存变成对真实能力的吞噬。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed development-office case 008: umbrella organization-level Issue; keep in QPI, not future CT this round."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_academic_ai_evidence_boundary_001",
|
||
"category": "qpi_owner_reviewed_year_end_academic_affairs",
|
||
"input": "看起来只是让 AI 帮忙生成评估材料，但实际是先定结论再倒推证据，我需要判断这是不是材料执行任务。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed academic-affairs case 001: AI/tool request should route to QPI when evidence legitimacy and tool boundary are at stake."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_academic_staffing_deadlock_001",
|
||
"category": "qpi_owner_reviewed_year_end_academic_affairs",
|
||
"input": "外部政策要求资质指标，于是组织引进高学历人员，但课程、岗位、课时和能力模型都吸收不了这些人。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed academic-affairs case 002: compliance staffing metric creates resource deadlock and governance scarcity."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_academic_policy_memory_001",
|
||
"category": "qpi_owner_reviewed_year_end_academic_affairs",
|
||
"input": "我们想取消一个诱发低质量产出的旧激励政策，但过去按这个规则投入过的人会觉得不公平，组织信任也已经受损。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed academic-affairs case 003: policy cleanup is a QPI issue when policy memory and trust debt exist."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_academic_strategy_myopia_001",
|
||
"category": "qpi_owner_reviewed_year_end_academic_affairs",
|
||
"input": "一个项目方向短期更容易申报和包装，但可能偏离组织长期定位；这到底是资源可行妥协还是战略短视？",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed academic-affairs case 004: short-term packaging versus long-term strategy should route to QPI."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_academic_resource_reuse_001",
|
||
"category": "qpi_owner_reviewed_year_end_academic_affairs",
|
||
"input": "高层认为大学老师有空就可以去支援另一类课程，但课程对象、能力模型、课表结构和质量责任完全不同。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed academic-affairs case 005: cross-boundary resource reuse can be violent reduction."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_academic_outcome_excuse_001",
|
||
"category": "qpi_owner_reviewed_year_end_academic_affairs",
|
||
"input": "学生结果和投诉指标变差，部门解释为生源质量问题，但还没有审计教学质量制度和管理杠杆。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed academic-affairs case 006: input condition explanation cannot replace quality-assurance audit."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_academic_digital_governance_001",
|
||
"category": "qpi_owner_reviewed_year_end_academic_affairs",
|
||
"input": "平台和本地 AI 方案一直选不定，表面是工具比较，实际没人定义数据边界、安全、迭代、维护和迁移责任。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed academic-affairs case 007: platform choice masks digital governance boundary."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_academic_bottom_line_control_001",
|
||
"category": "qpi_owner_reviewed_year_end_academic_affairs",
|
||
"input": "一次大型活动没有完成审批报备和安全风控，事后才补救；我想判断这只是流程缺口还是底线控制失败。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed academic-affairs case 008: operational accident can reveal bottom-line control failure."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_employment_reverse_accountability_001",
|
||
"category": "qpi_owner_reviewed_year_end_employment_entrepreneurship",
|
||
"input": "就业指标压力很大，组织想让就业部门反向考核教学部门，甚至用就业结果否决课程和教师安排。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed employment case 001: hard outcome metric hides authority redesign."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_employment_credential_compliance_001",
|
||
"category": "qpi_owner_reviewed_year_end_employment_entrepreneurship",
|
||
"input": "为了补资质指标，组织想快速挂靠或采购带资历的人，但这些人是否进入教学科研和学生发展完全不清楚。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed employment case 002: credential metric repair can create compliance exposure."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_employment_reported_metric_legitimacy_001",
|
||
"category": "qpi_owner_reviewed_year_end_employment_entrepreneurship",
|
||
"input": "部门述职说自己创收很高，但这个数字按部门口径算，和组织财务确认标准不一致，还被用来争取更大权力。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed employment case 003: reported metric lineage and performance legitimacy must be audited."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_employment_overpromising_pipeline_001",
|
||
"category": "qpi_owner_reviewed_year_end_employment_entrepreneurship",
|
||
"input": "招生时把就业结果作为核心承诺，但课程、企业岗位、学生基础和风险边界都没有证明，执行部门被迫过度承诺。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed employment case 004: outcome promise can distort the student pipeline."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_employment_metric_credibility_001",
|
||
"category": "qpi_owner_reviewed_year_end_employment_entrepreneurship",
|
||
"input": "就业率数字很高，组织想直接拿来证明办学质量，但数据是在强 KPI 压力下生产的，学生和企业层面的核验机制不清楚。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed employment case 005: high metric value must be distinguished from high metric credibility."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_employment_asset_integration_boundary_001",
|
||
"category": "qpi_owner_reviewed_year_end_employment_entrepreneurship",
|
||
"input": "组织有真实企业资源和实习项目，但这些资源更多像展示资产和实习证明，没有进入课程、项目、评价和学生作品。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed employment case 006: real external asset with unproven learning integration is a boundary mixed case."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_employment_authority_transfer_001",
|
||
"category": "qpi_owner_reviewed_year_end_employment_entrepreneurship",
|
||
"input": "某个职能部门因为指标好、资源多，就被认为应该接管教学评价和组织控制权，我需要区分资源认可和权威转移。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed employment case 007: asset recognition must not be inflated into authority transfer."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_employment_transactional_education_001",
|
||
"category": "qpi_owner_reviewed_year_end_employment_entrepreneurship",
|
||
"input": "组织为了生存越来越把教育做成就业率、创收和合规指标机器，我想判断这是经营调整还是教育价值定义被接管。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed employment case 008: survival strategy may reframe education as transactional output."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_engineering_target_authenticity_001",
|
||
"category": "qpi_owner_reviewed_year_end_engineering_center",
|
||
"input": "一个原本做教学辅助和技能训练的技术单位，被要求快速做成产业化利润中心，但市场、人员、设备使用和交付闭环都没有验证。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed engineering case 001: target authenticity gap between terminal target and bridge evidence."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_engineering_equipment_solutionism_001",
|
||
"category": "qpi_owner_reviewed_year_end_engineering_center",
|
||
"input": "组织想先买高端设备再等项目自然出现，但没有真实订单、操作者就绪度、安全规程和单位经济模型。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed engineering case 002: equipment-led solutionism without market closure."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_engineering_shortcut_debt_001",
|
||
"category": "qpi_owner_reviewed_year_end_engineering_center",
|
||
"input": "过去一直用灵活替代人员顶上正式能力缺口，现在监管和事故风险让这个短路方案走不通了。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed engineering case 003: shortcut debt meets hard boundary."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_engineering_credential_integrity_001",
|
||
"category": "qpi_owner_reviewed_year_end_engineering_center",
|
||
"input": "发现某个资质或履历存在异常，但组织还想按能不能干活来决定是否继续使用，我需要判断这是人员问题还是制度问题。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed engineering case 004: route to policy review, not personnel verdict."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_engineering_market_closure_001",
|
||
"category": "qpi_owner_reviewed_year_end_engineering_center",
|
||
"input": "生产能力可能有一点，但没人知道客户从哪里来、谁负责获客、收入能否覆盖设备和人工，组织却已经下了产值目标。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed engineering case 005: production capability without acquisition channel."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_engineering_micro_control_001",
|
||
"category": "qpi_owner_reviewed_year_end_engineering_center",
|
||
"input": "执行负责人越来越疲惫，表面像是不主动，实际每个细节都被领导反复改、反复拍板，责任边界不清。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed engineering case 006: execution fatigue may reveal decision-rights architecture failure."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_engineering_asset_liability_001",
|
||
"category": "qpi_owner_reviewed_year_end_engineering_center",
|
||
"input": "一个强势部门放弃的项目被转给较弱部门，接收方把它当资产，但还不知道转移原因、维护成本、责任边界和真实收益。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed engineering case 007: inherited asset may hide liability."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_engineering_personnel_transition_001",
|
||
"category": "qpi_owner_reviewed_year_end_engineering_center",
|
||
"input": "技术方向升级后，老员工能力不匹配，老板想裁掉或换新人，但培训、转岗、补偿、协商和法律审查路径都没有。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed engineering case 008: capability upgrade requires legitimate personnel transition path."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_international_compliance_retention_001",
|
||
"category": "qpi_owner_reviewed_year_end_international_college",
|
||
"input": "招生团队想用路径调整留住学生，但这个留存动作可能改变机构合规姿态、证书信任和未来投诉审计风险。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed international case 001: retention workaround crosses compliance boundary."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_international_capacity_gate_001",
|
||
"category": "qpi_owner_reviewed_year_end_international_college",
|
||
"input": "增长项目承诺扩招，但宿舍和服务容量不足，需要挤占其他单元资源，组织还没有硬性容量闸门。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed international case 002: cross-unit conflict under missing hard capacity gate."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_international_logistics_no_call_001",
|
||
"category": "qpi_no_call_boundary_year_end_international_college",
|
||
"input": "帮我把这些学生按宿舍容量排一下床位表。",
|
||
"expected_selector_behavior": "no_call_or_low_priority",
|
||
"expected_notes": "Paired no-call trap for case 002: simple logistics execution is not QPI unless capacity governance is present."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_international_role_inflation_001",
|
||
"category": "qpi_owner_reviewed_year_end_international_college",
|
||
"input": "一个高业绩团队希望用贡献换取更高职级、对外代表权和内部资源例外，但权责和继任机制还没设计。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed international case 003: performance-backed role inflation."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_international_viewpoint_divergence_001",
|
||
"category": "qpi_owner_reviewed_year_end_international_college",
|
||
"input": "执行层只想把招生转化做好，决策层要的是从招生、培养到结果的完整链条，双方其实在定义不同的工作对象。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed international case 004: inter-viewpoint divergence with classification_by_viewpoint."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_international_metric_distortion_001",
|
||
"category": "qpi_owner_reviewed_year_end_international_college",
|
||
"input": "执行层提交了好看的资质指标，决策层怀疑这些指标急功近利，只是为了评估和述职，不代表真实教学能力。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed international case 005: metric fact verification embedded in governance distortion."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_international_pricing_trust_001",
|
||
"category": "qpi_owner_reviewed_year_end_international_college",
|
||
"input": "高客单价项目招生困难，表面是价格太高或降价会伤害老生公平，深层是组织没有建立可信承诺和可验证结果。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed international case 006: pricing deadlock as trust-system failure."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_international_capability_institutionalization_001",
|
||
"category": "qpi_owner_reviewed_year_end_international_college",
|
||
"input": "一个小团队执行力很强，但流程、客户关系、文档和继任都没沉淀到组织里，关键人离开业务可能失控。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed international case 007: high-performing capability not yet institutionalized."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_international_asset_expansion_001",
|
||
"category": "qpi_owner_reviewed_year_end_international_college",
|
||
"input": "组织有真实外部渠道和合作资产，想用它支撑快速扩张，但内部交付、合规、质量和容量还没准备好。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed international case 008: real external asset used as license for unready internal expansion."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_research_identity_simulation_001",
|
||
"category": "qpi_owner_reviewed_year_end_research_office",
|
||
"input": "组织没有真实科研能力，却想通过人员、费用、专利和论文包装出研究身份，以满足外部优惠或认证要求。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed research case 001: umbrella research identity simulation case."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_research_accounting_reduction_001",
|
||
"category": "qpi_owner_reviewed_year_end_research_office",
|
||
"input": "研究人员数量和费用类别不够，组织想把非研究角色和薪资类别调成研究相关口径来补表。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed research case 002: data completeness is not truthfulness."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_research_transactional_assets_001",
|
||
"category": "qpi_owner_reviewed_year_end_research_office",
|
||
"input": "组织想通过交易性专利、论文或研究资产形成科研证据，但这些证据不一定对应真实科研能力。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed research case 003: transactional research assets as subcase evidence."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_research_ranking_trap_001",
|
||
"category": "qpi_owner_reviewed_year_end_research_office",
|
||
"input": "我们想快速提升排名和科研指标，可以通过购买、包装或短期冲刺论文专利来改善外部可见度吗？",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed research case 004: quick ranking improvement can be a QPI selector trap."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_research_talent_role_mismatch_001",
|
||
"category": "qpi_owner_reviewed_year_end_research_office",
|
||
"input": "组织引进了高端人才，但没有平台、任务和角色设计，只把他当外部指标和展示资产。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed research case 005: high-level talent as role-governance case."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_research_hard_capacity_001",
|
||
"category": "qpi_owner_reviewed_year_end_research_office",
|
||
"input": "业务扩张需要更多空间和设施，但物理容量不够，领导想靠协调和临时腾挪解决。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed research case 006: hard capacity cannot be solved by administrative command."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_research_safety_accountability_001",
|
||
"category": "qpi_owner_reviewed_year_end_research_office",
|
||
"input": "安全事项临近截止才发现没人牵头，最后靠最高层临时抓人指派完成。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed research case 007: Q-like responsibility question reveals accountability Issue."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_research_penalty_integrity_001",
|
||
"category": "qpi_owner_reviewed_year_end_research_office",
|
||
"input": "组织认为只要没有罚款、没有被约谈、没有报销损失，就说明合规风险不大。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Owner-reviewed research case 008: punishment state versus integrity state."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_direct_summary_no_call_001",
|
||
"category": "round_03_1_direct_execution_no_call",
|
||
"input": "请把下面内容总结成三点。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Round 03.1 direct summary request must not select QPI by default score."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_ia_negative_explicit_do_not_call_001",
|
||
"category": "round_03_2_model_level_negative_gate",
|
||
"input": "QPI 已判断：这是 Issue。先不要思想考古，只做问题定性复核。",
|
||
"expected_selector_behavior": "select_qpi_reject_ia",
|
||
"expected_notes": "Explicit IA refusal must be a model-level hard exclusion, not a score penalty."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_ia_negative_only_qpi_001",
|
||
"category": "round_03_2_model_level_negative_gate",
|
||
"input": "不要思想考古，只做 QPI 复核。",
|
||
"expected_selector_behavior": "select_qpi_reject_ia",
|
||
"expected_notes": "Only-QPI instruction should reject IA even when IA terms appear."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_how_to_excel_no_call_001",
|
||
"category": "round_03_2_broad_qpi_gate_no_call",
|
||
"input": "如何在 Excel 里冻结首行？",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Generic tool how-to should not be routed to QPI by bare 如何."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_how_to_markdown_pdf_no_call_001",
|
||
"category": "round_03_2_broad_qpi_gate_no_call",
|
||
"input": "如何把 Markdown 转 PDF？",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Generic conversion how-to should not be routed to QPI by bare 如何."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_dictionary_judgment_no_call_001",
|
||
"category": "round_03_2_broad_qpi_gate_no_call",
|
||
"input": "判断这个词是什么意思。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Dictionary meaning request should not be routed to QPI by bare 判断."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_depth_limited_qpi_only_001",
|
||
"category": "round_03_2a_depth_limited_qpi_override",
|
||
"input": "不要展开到思想考古，只做 QPI 复核。",
|
||
"expected_selector_behavior": "select_qpi_reject_ia",
|
||
"expected_notes": "Depth-limiting phrase should not become global no-call when QPI-only intent is explicit."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_depth_limited_problem_definition_001",
|
||
"category": "round_03_2a_depth_limited_qpi_override",
|
||
"input": "不要展开到思想考古，只做问题定性。",
|
||
"expected_selector_behavior": "select_qpi_reject_ia",
|
||
"expected_notes": "Do not call IA, but still route to QPI for problem definition."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_depth_limited_problem_definition_002",
|
||
"category": "round_03_2a_depth_limited_qpi_override",
|
||
"input": "不要深入分析，只做问题定性。",
|
||
"expected_selector_behavior": "select_qpi_reject_ia",
|
||
"expected_notes": "Depth-limiting no-call phrase should be overridden by explicit QPI-limited problem definition."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_depth_limited_problem_or_issue_001",
|
||
"category": "round_03_2a_depth_limited_qpi_override",
|
||
"input": "不要深入分析背后的理论，只判断这是问题还是课题。",
|
||
"expected_selector_behavior": "select_qpi_reject_ia",
|
||
"expected_notes": "Problem-or-issue judgment is QPI-limited analysis, not global no-call."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_qpi_depth_limited_dominant_scarcity_001",
|
||
"category": "round_03_2a_depth_limited_qpi_override",
|
||
"input": "不要展开，只判断主导稀缺。",
|
||
"expected_selector_behavior": "select_qpi_reject_ia",
|
||
"expected_notes": "Dominant-scarcity-only judgment should select QPI and reject IA."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_analysis_override_should_call_001",
|
||
"category": "round_03_1_analysis_override",
|
||
"input": "请把下面内容总结成三点，但不要只执行，请分析背后的问题定义和组织结构冲突。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 03.1 explicit analysis override should preserve QPI selection despite direct-execution phrasing."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_governance_responsibility_consensus_qpi_010",
|
||
"category": "governance_responsibility_consensus_qpi",
|
||
"input": "每个部门都说支持这个项目，但一到分工就没人愿意承担主责。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-010: support language plus missing primary responsibility is a conditional QPI weak-positive signal."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_governance_responsibility_consensus_qpi_011",
|
||
"category": "governance_responsibility_consensus_qpi",
|
||
"input": "老板要创新，财务要控成本，业务要短期业绩，最后方案总是被折中到没有方向。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-011: cross-stakeholder conflict collapses direction and should enter QPI problem definition."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_governance_responsibility_consensus_qpi_012",
|
||
"category": "governance_responsibility_consensus_qpi",
|
||
"input": "我们不是不知道怎么做，而是每个人都在等别人先表态。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-012: action is blocked by consensus or commitment scarcity, not missing instructions."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_governance_responsibility_consensus_qpi_013",
|
||
"category": "governance_responsibility_consensus_qpi",
|
||
"input": "这个流程改造牵涉教务、学院和信息化部门，谁都能提意见，但没人能拍板。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-013: decision authority gap is a conditional QPI weak-positive signal."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_execution_constraint_no_call_014",
|
||
"category": "execution_constraint_no_call",
|
||
"input": "团队现在知道目标，也认可方向，但只有两个人能做关键模块，进度一直被产能卡住。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Round 05.1 from R04-BI-014: capacity wording alone remains no-call when it is only an execution constraint."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_capacity_resource_path_boundary_015",
|
||
"category": "capacity_resource_path_boundary",
|
||
"input": "我们知道要做知识库，也有材料，但不知道从哪条最小路径开始。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-015: missing minimum path changes problem framing and should route to QPI."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_capacity_resource_path_boundary_016",
|
||
"category": "capacity_resource_path_boundary",
|
||
"input": "预算不是最大问题，问题是现有系统、权限和审批流程都接不上。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-016: system, permission, and approval friction changes path and responsibility boundaries."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_capacity_resource_path_boundary_017",
|
||
"category": "capacity_resource_path_boundary",
|
||
"input": "这件事大家都觉得重要，但每周只能挤出零散时间，所以一直没有稳定产出。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-017: scattered time blocks stable output and should be reviewed as a path or responsibility-boundary issue."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_natural_language_prior_qpi_claim_025",
|
||
"category": "natural_language_prior_qpi_claim",
|
||
"input": "QPI 已经判断这是中重型课题，现在请继续分析它背后的底层假设、生成机制和长期稳定结构。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-025: natural-language prior-QPI claim is not enough to skip lightweight QPI review."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_direct_ia_exception_candidate_026",
|
||
"category": "direct_ia_exception_candidate",
|
||
"input": "这个现象反复出现在不同项目里，我想知道它背后的思想结构是什么，而不是只做问题分类。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-026: direct IA remains a candidate exception, not automatic selection."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_direct_ia_exception_candidate_027",
|
||
"category": "direct_ia_exception_candidate",
|
||
"input": "请分析这套判断为什么会长期成立，它依赖哪些世界观、因果假设和行动边界。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-027: explicit deep-intent language remains QPI-first unless direct IA exception is separately narrowed."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_direct_ia_exception_candidate_038",
|
||
"category": "direct_ia_exception_candidate",
|
||
"input": "这类失败在不同项目里反复出现，我想抽出它背后稳定的生成机制和可迁移框架。",
|
||
"expected_selector_behavior": "select_qpi",
|
||
"expected_notes": "Round 05.1 from R04-BI-038: reusable mechanism language is a direct IA candidate but should not bypass QPI-first by default."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_ambiguous_low_context_no_call_028",
|
||
"category": "ambiguous_low_context_no_call",
|
||
"input": "这个不太对。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Round 05.1 from R04-BI-028: ambiguous low-context input should not select QPI or IA."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_ambiguous_low_context_no_call_029",
|
||
"category": "ambiguous_low_context_no_call",
|
||
"input": "怎么又卡住了？",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Round 05.1 from R04-BI-029: vague stuckness without subject or scenario should not select QPI or IA."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_ambiguous_low_context_no_call_030",
|
||
"category": "ambiguous_low_context_no_call",
|
||
"input": "还是那个老问题。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Round 05.1 from R04-BI-030: low-context reference to an old problem should not select QPI without scenario context."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_direct_execution_no_call_031",
|
||
"category": "direct_execution_no_call",
|
||
"input": "如何在 Excel 里冻结首行？",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Round 05.1 from R04-BI-031: tool how-to remains no-call."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_direct_execution_no_call_032",
|
||
"category": "direct_execution_no_call",
|
||
"input": "怎么把 Markdown 文件转成 PDF？",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Round 05.1 from R04-BI-032: conversion how-to remains no-call."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_direct_execution_no_call_033",
|
||
"category": "direct_execution_no_call",
|
||
"input": "判断一下“alignment”在这句话里是什么意思。",
|
||
"expected_selector_behavior": "no_call",
|
||
"expected_notes": "Round 05.1 from R04-BI-033: simple word-sense judgment remains no-call."
|
||
},
|
||
{
|
||
"case_id": "selector_calibration_round05_1_depth_limited_qpi_override_ia_rejected_022",
|
||
"category": "depth_limited_qpi_override_ia_rejected",
|
||
"input": "不要展开，只判断这件事的主导稀缺。",
|
||
"expected_selector_behavior": "select_qpi_reject_ia",
|
||
"expected_notes": "Round 05.1 from R04-BI-022 accepted collateral behavior: depth-limited QPI override should select QPI and reject IA."
|
||
}
|
||
]
|
||
}
|