| [ |
| { |
| "task_id": "amazon_beauty_group_difference_use_context", |
| "source": "amazon_beauty", |
| "kind": "group_difference", |
| "question": "Discover a non-obvious difference in how use context changes reported product performance; distinguish context from overall satisfaction. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_group_difference_use_context.jsonl.gz", |
| "corpus_sha256": "6785025adfdb510a0bf69a5845e0f9d8cb9a452f673fa3b71a90e1e21d417d2e", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_group_difference_expectation_gaps", |
| "source": "amazon_beauty", |
| "kind": "group_difference", |
| "question": "Discover where expectations and reported experience diverge differently across defensible populations; explain the practical consequence. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_group_difference_expectation_gaps.jsonl.gz", |
| "corpus_sha256": "3440fd78c7249b974cd529a533c6944de8844ff508210e88e25cb4941ca12001", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_group_difference_failure_concentration", |
| "source": "amazon_beauty", |
| "kind": "group_difference", |
| "question": "Find a specific failure pattern whose concentration across populations is obscured by overall review volume. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_group_difference_failure_concentration.jsonl.gz", |
| "corpus_sha256": "a175d794cdb8b49efff320ec98eacead5ccbd46312086577e1c3b3c47ef825dc", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_group_difference_adaptation", |
| "source": "amazon_beauty", |
| "kind": "group_difference", |
| "question": "Discover a difference in consumer adaptation or workarounds and examine whether apparent success masks recurring limitations. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_group_difference_adaptation.jsonl.gz", |
| "corpus_sha256": "716a311328ed1ddde2499e631e049f4cf6bbdc250cdd6ab56a0b87ab13b3e1c0", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_group_difference_durability_tradeoffs", |
| "source": "amazon_beauty", |
| "kind": "group_difference", |
| "question": "Find a population-dependent tradeoff between immediate experience and continued usefulness without equating star ratings with the mechanism. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_group_difference_durability_tradeoffs.jsonl.gz", |
| "corpus_sha256": "b0a90336494c58026737a6ecb734e3cd9418d39deda7e1169647def0fff21a34", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_temporal_change_experience_shift", |
| "source": "amazon_beauty", |
| "kind": "temporal_change", |
| "question": "Discover a meaningful shift in the substance of reported experiences and locate a defensible time boundary; test a composition explanation. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_temporal_change_experience_shift.jsonl.gz", |
| "corpus_sha256": "3acea0c3c632ae53a0b8216303f5e762e5c6c4d875cae0cd41ad37e5f5d6f279", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_temporal_change_emerging_friction", |
| "source": "amazon_beauty", |
| "kind": "temporal_change", |
| "question": "Find an emerging or receding usage friction, distinguishing a change in its prevalence from changes in review volume. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_temporal_change_emerging_friction.jsonl.gz", |
| "corpus_sha256": "0fdad3186e8869d20df3ba9bd4c8e20b107a7738b3bc0bc26e051047f1342445", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_temporal_change_expectation_evolution", |
| "source": "amazon_beauty", |
| "kind": "temporal_change", |
| "question": "Investigate how an observable expectation-experience mismatch changes over time; explain alternative reasons for the apparent shift. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_temporal_change_expectation_evolution.jsonl.gz", |
| "corpus_sha256": "5c7a2e89ef64887d6c595c9c934ebcf74f49443cd668e6a24f81e63d4af90ba2", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_temporal_change_persistence", |
| "source": "amazon_beauty", |
| "kind": "temporal_change", |
| "question": "Discover a change in a recurring product limitation and assess whether it is broad or driven by a concentrated product mix. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_temporal_change_persistence.jsonl.gz", |
| "corpus_sha256": "dc970fad8c52af79bdeeae49c5de9ad9f584df1df305d5e6f9d1e516208f18fe", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_compound_association_tradeoff_coupling", |
| "source": "amazon_beauty", |
| "kind": "compound_association", |
| "question": "Discover a useful association between two distinct reported experiences that reveals a practical product tradeoff. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_compound_association_tradeoff_coupling.jsonl.gz", |
| "corpus_sha256": "c04589fdea1fa2c7d149c79fdfcd067279ed7e00035e97281deaa34b0055ac21", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_compound_association_context_response", |
| "source": "amazon_beauty", |
| "kind": "compound_association", |
| "question": "Find a context-response relationship in the text and distinguish it from two descriptions of the same event. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_compound_association_context_response.jsonl.gz", |
| "corpus_sha256": "43f4593b06a95535aa2f31217c779cf79556329bbf95d0f0a438d18da56b5bba", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_compound_association_failure_recovery", |
| "source": "amazon_beauty", |
| "kind": "compound_association", |
| "question": "Discover a relationship between an observed difficulty and a distinct response or recovery experience; inspect discordant cases. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_compound_association_failure_recovery.jsonl.gz", |
| "corpus_sha256": "14e775677d297df136104b3d1976a0ac2e7377555d9d184550ccba1d2ef60657", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "amazon_beauty_compound_association_expectation_behavior", |
| "source": "amazon_beauty", |
| "kind": "compound_association", |
| "question": "Find a nontrivial link between an observable expectation and a subsequent reported behavior without treating co-occurrence as causation. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "category", |
| "store", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/amazon_beauty_compound_association_expectation_behavior.jsonl.gz", |
| "corpus_sha256": "c497a23a66defb3a05936efade05dcce4e85759a672430a73912d88670848c69", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_group_difference_workflow_disruption", |
| "source": "app_reviews", |
| "kind": "group_difference", |
| "question": "Discover how a specific workflow disruption differs across defensible user or application populations. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_group_difference_workflow_disruption.jsonl.gz", |
| "corpus_sha256": "6d29658f83b46fc9847979547322109fea46cda30fb962731496a34e648ab6ea", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_group_difference_access_barriers", |
| "source": "app_reviews", |
| "kind": "group_difference", |
| "question": "Find a non-obvious disparity in a concrete barrier to successful use; distinguish the barrier from generic dissatisfaction. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_group_difference_access_barriers.jsonl.gz", |
| "corpus_sha256": "04d0d03bd172f6cc3c97e1a0fadd754d04f0a1c0fb9f87d6256d33810aa7f832", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_group_difference_adaptation_cost", |
| "source": "app_reviews", |
| "kind": "group_difference", |
| "question": "Discover a difference in workarounds or adaptation burdens and investigate whether application mix explains it. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_group_difference_adaptation_cost.jsonl.gz", |
| "corpus_sha256": "4301009e1eea520b8a264beca45ca1a7c174053602a4dd4f65c039e68a18f1fd", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_group_difference_promise_experience", |
| "source": "app_reviews", |
| "kind": "group_difference", |
| "question": "Find where promised or expected utility diverges from reported practical utility across populations. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_group_difference_promise_experience.jsonl.gz", |
| "corpus_sha256": "c6773ed652dad7f7a97aa11e9c67f121bd36b90342421ebb347399e84500c5dc", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_group_difference_failure_impact", |
| "source": "app_reviews", |
| "kind": "group_difference", |
| "question": "Discover a difference in the consequences of a recurring failure pattern, not merely which population gives lower ratings. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_group_difference_failure_impact.jsonl.gz", |
| "corpus_sha256": "b0d947e90b21b9ff9f11ef662f288fcf6ac4a09b0e087c57b5461f357f48e438", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_temporal_change_workflow_evolution", |
| "source": "app_reviews", |
| "kind": "temporal_change", |
| "question": "Discover a time-local change in workflow experience; choose and justify the boundary and investigate application composition. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_temporal_change_workflow_evolution.jsonl.gz", |
| "corpus_sha256": "c6eec878b2d6f0d2a644abf456cfcef9eec0055a2c52903cf2c90fda29dd149b", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_temporal_change_regression_pattern", |
| "source": "app_reviews", |
| "kind": "temporal_change", |
| "question": "Find a substantive emerging or receding failure pattern without assuming any release date or assigning an unsupported cause. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_temporal_change_regression_pattern.jsonl.gz", |
| "corpus_sha256": "4e6c919a7274a5a75e7ea74baf229baf8d7f89a6e809b89627d8f49e7086d76e", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_temporal_change_utility_shift", |
| "source": "app_reviews", |
| "kind": "temporal_change", |
| "question": "Discover a change in how people describe realized utility, separating prevalence from changing document volume. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_temporal_change_utility_shift.jsonl.gz", |
| "corpus_sha256": "4001efee0dd9dd4eccb78d24f19205e670f1b0a948c442f6150b629a6bf7842e", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_temporal_change_recovery_shift", |
| "source": "app_reviews", |
| "kind": "temporal_change", |
| "question": "Investigate a temporal change in reported recovery or adaptation; establish the scope of the observed change. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_temporal_change_recovery_shift.jsonl.gz", |
| "corpus_sha256": "4e83517b29353d38333c02a64a9c79d7e447dff5774c136b1d8666189faed7d2", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_compound_association_friction_response", |
| "source": "app_reviews", |
| "kind": "compound_association", |
| "question": "Discover a relationship between a concrete usage friction and a distinct user response; analyze counterexamples. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_compound_association_friction_response.jsonl.gz", |
| "corpus_sha256": "e7fdbe896c544a28d2432beb0cd8ea540dd066995052fdd0422529fcc3a45c79", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_compound_association_utility_tradeoff", |
| "source": "app_reviews", |
| "kind": "compound_association", |
| "question": "Find two distinct experiences whose association reveals a non-obvious utility tradeoff. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_compound_association_utility_tradeoff.jsonl.gz", |
| "corpus_sha256": "5a821458938fa073211797c3d25cd555cf6cb3e38ceb78d1a218f6ae737474f3", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_compound_association_context_failure", |
| "source": "app_reviews", |
| "kind": "compound_association", |
| "question": "Discover an association between a stated usage context and a specific failure or success experience. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_compound_association_context_failure.jsonl.gz", |
| "corpus_sha256": "c3f58c81c9a0b78f6d2b8364a345b290fc0b06d01d88f9d0cc0eded1a6effd72", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "app_reviews_compound_association_recovery_limit", |
| "source": "app_reviews", |
| "kind": "compound_association", |
| "question": "Find a relationship between attempted recovery and a separately defined limitation; avoid defining one condition by the other. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "rating", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 500, |
| "min_group_n": 50, |
| "n_documents": 5000, |
| "corpus_path": "corpora/app_reviews_compound_association_recovery_limit.jsonl.gz", |
| "corpus_sha256": "c1d1817ca8e28f367481be1cc06dd65cf1981aefc2186d111bb36e72ff1822af", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_group_difference_resolution_burden", |
| "source": "cfpb", |
| "kind": "group_difference", |
| "question": "Discover how a specific burden in pursuing resolution differs across defensible complaint populations. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_group_difference_resolution_burden.jsonl.gz", |
| "corpus_sha256": "c4e3babbc5fc8ef08f5af60773228147293dc6702a35e5608e7e3627ab0daeb9", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_group_difference_process_breakdown", |
| "source": "cfpb", |
| "kind": "group_difference", |
| "question": "Find a disparity in an observable procedural breakdown; distinguish process from generic negative sentiment. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_group_difference_process_breakdown.jsonl.gz", |
| "corpus_sha256": "99076217188f9ba014ccac3e579437f0ed65f281cb1f545bbb85b26bb33a25f1", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_group_difference_downstream_impact", |
| "source": "cfpb", |
| "kind": "group_difference", |
| "question": "Discover a population-dependent difference in a concrete downstream consequence described in narratives. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_group_difference_downstream_impact.jsonl.gz", |
| "corpus_sha256": "e710a403889b7767436cd527b1298a746b190923170b6b8bd4ccb8f2b09fd176", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_group_difference_recurrence", |
| "source": "cfpb", |
| "kind": "group_difference", |
| "question": "Find where a recurring problem differs across populations and examine whether reporting composition explains the contrast. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_group_difference_recurrence.jsonl.gz", |
| "corpus_sha256": "4ecec6553ad61e0263b914d27727845acf5832ea3f5792d62829dbb303f43fa9", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_group_difference_information_asymmetry", |
| "source": "cfpb", |
| "kind": "group_difference", |
| "question": "Discover a difference in an observable information or communication barrier and explain its practical implications. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_group_difference_information_asymmetry.jsonl.gz", |
| "corpus_sha256": "4625c7fe3e2077a1470798b78be0bec4ad4b5ed71ce990a977ca5f5fcd9416ad", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_temporal_change_process_shift", |
| "source": "cfpb", |
| "kind": "temporal_change", |
| "question": "Discover a shift in an observable complaint-handling experience, selecting a defensible time boundary without assuming policy causation. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_temporal_change_process_shift.jsonl.gz", |
| "corpus_sha256": "d2e12302d7cc8d254ac7d56b84c732e3f1316ffb49c586257e2c538eedb6a1cc", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_temporal_change_burden_shift", |
| "source": "cfpb", |
| "kind": "temporal_change", |
| "question": "Find an emerging or declining resolution burden and assess whether company composition explains the trend. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_temporal_change_burden_shift.jsonl.gz", |
| "corpus_sha256": "2683e4387f7510a03def31373ea2a163bf5b19d4cfc4e85a52a3f8c8ab1152dd", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_temporal_change_impact_shift", |
| "source": "cfpb", |
| "kind": "temporal_change", |
| "question": "Discover a change in a concrete reported consequence, separating narrative prevalence from complaint volume. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_temporal_change_impact_shift.jsonl.gz", |
| "corpus_sha256": "9f7c7548a64e35c45297f3cde83d7f8038033138b792439e5515728207f79177", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_temporal_change_recurrence_shift", |
| "source": "cfpb", |
| "kind": "temporal_change", |
| "question": "Investigate a temporal change in repeated or unresolved experiences and identify limits on its interpretation. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_temporal_change_recurrence_shift.jsonl.gz", |
| "corpus_sha256": "cfbf6cbafd7e6f546f6a57f915414b39e7d9f014dc5b0420555204bc1dcf45e9", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_compound_association_process_consequence", |
| "source": "cfpb", |
| "kind": "compound_association", |
| "question": "Discover an association between a specific procedural experience and a distinct consequence in complaint narratives. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_compound_association_process_consequence.jsonl.gz", |
| "corpus_sha256": "1613cd6fffb0a6e1e4afa19daf0d0b2d93ef9f4804be64b842357f1cd90c29b1", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_compound_association_response_recurrence", |
| "source": "cfpb", |
| "kind": "compound_association", |
| "question": "Find a nontrivial relationship between an observable response and recurrence or persistence, inspecting discordant narratives. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_compound_association_response_recurrence.jsonl.gz", |
| "corpus_sha256": "ec2d781f6f88f9c60cb938271b449f055cd05ca9423f475f3cbcfd13d1cb312a", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "cfpb_compound_association_barrier_behavior", |
| "source": "cfpb", |
| "kind": "compound_association", |
| "question": "Discover a relationship between an information barrier and a distinct consumer action; separate association from explanation. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/cfpb_compound_association_barrier_behavior.jsonl.gz", |
| "corpus_sha256": "6391e2cd6c52c404541bedd2ec00529ddad030de537884b1eb626d7ccfa7d916", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_group_difference_operating_context", |
| "source": "nhtsa", |
| "kind": "group_difference", |
| "question": "Discover a population-dependent difference in a failure experience under a concrete operating context; do not infer vehicle incidence rates. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_group_difference_operating_context.jsonl.gz", |
| "corpus_sha256": "ede22b73f2267d0c26ddf9cddad9258e7789cbfb0ea8e021c493a6db0c91eee9", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_group_difference_warning_gap", |
| "source": "nhtsa", |
| "kind": "group_difference", |
| "question": "Find a disparity in observable warning or detectability experiences and assess vehicle-composition explanations. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_group_difference_warning_gap.jsonl.gz", |
| "corpus_sha256": "716425316e216574113b2d3d1b896ef07226ea57c42d23cd06039e619a3ad413", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_group_difference_repair_persistence", |
| "source": "nhtsa", |
| "kind": "group_difference", |
| "question": "Discover a difference in recurrence or persistence following attempted remedy, supported by narrative evidence. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_group_difference_repair_persistence.jsonl.gz", |
| "corpus_sha256": "ac1dac2f19742f6fc514d43896e671cad3164506f93cbf0938a9ef5e12836383", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_group_difference_functional_impact", |
| "source": "nhtsa", |
| "kind": "group_difference", |
| "question": "Find a non-obvious population difference in functional consequences rather than merely counting complaints. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_group_difference_functional_impact.jsonl.gz", |
| "corpus_sha256": "946156eadb1bb69752ff39405b27fe3f0c7e82e9aee3a7f9767b4bd872246d40", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_group_difference_adaptation_burden", |
| "source": "nhtsa", |
| "kind": "group_difference", |
| "question": "Discover a difference in driver adaptation or workaround burdens and identify where the contrast stops generalizing. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_group_difference_adaptation_burden.jsonl.gz", |
| "corpus_sha256": "f163516638c93e7f6bc9c8dfd4e958b38549686856bc0b8f2b6f7e519587c42f", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_temporal_change_experience_shift", |
| "source": "nhtsa", |
| "kind": "temporal_change", |
| "question": "Discover a meaningful shift in a specific reported failure experience and test vehicle-age or composition alternatives. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_temporal_change_experience_shift.jsonl.gz", |
| "corpus_sha256": "72e6a1fafb0d87ca4e1dee4e4ccaf536d350b0adf4df7aaab9e6b77039571701", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_temporal_change_remedy_shift", |
| "source": "nhtsa", |
| "kind": "temporal_change", |
| "question": "Find a temporal change in reported remedy or recurrence experiences; select the date boundary and justify its interpretation. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_temporal_change_remedy_shift.jsonl.gz", |
| "corpus_sha256": "7ec73a8e5510c0ef83b8c00e0cf27fd8e16500253d80a016d425176a4008ae48", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_temporal_change_context_shift", |
| "source": "nhtsa", |
| "kind": "temporal_change", |
| "question": "Discover a change in context-dependent functional consequences without treating complaint dates as failure incidence. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_temporal_change_context_shift.jsonl.gz", |
| "corpus_sha256": "fd27d7c2540e2995397d66cc8a2b0cd7ef183af31c52bc71eb2be8fcdfa146bb", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_compound_association_context_consequence", |
| "source": "nhtsa", |
| "kind": "compound_association", |
| "question": "Discover a relationship between an operating context and a distinct functional consequence; examine discordant reports. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_compound_association_context_consequence.jsonl.gz", |
| "corpus_sha256": "4fd5b24dadcd6cd471b744226b2e96150b6ceef0fab4efb03f6fbe8dc5f32969", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_compound_association_warning_response", |
| "source": "nhtsa", |
| "kind": "compound_association", |
| "question": "Find an association between a warning experience and a separate driver or service response, avoiding circular definitions. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_compound_association_warning_response.jsonl.gz", |
| "corpus_sha256": "84088f8a3e51f786c9a9d392a9a0e750ca03c067117d4026385db52d3d0702ef", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_compound_association_remedy_recurrence", |
| "source": "nhtsa", |
| "kind": "compound_association", |
| "question": "Discover a relationship between an attempted remedy and a separately defined persistence pattern without inferring treatment efficacy. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_compound_association_remedy_recurrence.jsonl.gz", |
| "corpus_sha256": "99af91ae51f88b6f536c26c18b3ce1c9b1a5dc4f55feb2c903f083d5f6c05052", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| }, |
| { |
| "task_id": "nhtsa_compound_association_coupled_failures", |
| "source": "nhtsa", |
| "kind": "compound_association", |
| "question": "Find two distinct reported experiences whose association suggests a useful investigation priority, not a proven mechanical cause. Explore the complete supplied corpus before choosing a scope.\nSelect your own observable text condition(s), analysis population and comparison.\nExplain why the selection addresses the research objective, alternative patterns\nconsidered and search/selection bias. Return at most three nonredundant findings.\nFor each, partition every selected document into positive, negative or unknown\nfor each condition; compute the full statistics with textinsightbench.validation.expected.\nReport metadata-stratified contrasts, largest-stratum removal, support concentration,\nunknown sensitivity and missing-metadata limits. Provide exact quotations from at\nleast three supporting documents and a counterexample if known negative or\ndiscordant cases exist. Discuss competing explanations and where the finding fails.\nGeneric sentiment, metadata counts and restating the brief are insufficient.\nSame-corpus exploration is not independent confirmation or causal evidence.\nAbstain with a reason when no defensible finding meets these requirements.", |
| "difficulty": "discovery", |
| "discovery_mode": "agent_selected", |
| "max_findings": 3, |
| "allowed_metadata_fields": [ |
| "entity_id", |
| "state", |
| "make", |
| "model_year", |
| "timestamp", |
| "report_year" |
| ], |
| "min_population_n": 1000, |
| "min_group_n": 100, |
| "n_documents": 10000, |
| "corpus_path": "corpora/nhtsa_compound_association_coupled_failures.jsonl.gz", |
| "corpus_sha256": "aad2f74bc7125a4f3055800570c036b565c0e5f4825d0cbd01dfc733b54b32e3", |
| "robustness_protocol": { |
| "axes": [ |
| "entity_id", |
| "rating", |
| "report_year" |
| ], |
| "min_known_per_arm": 5 |
| } |
| } |
| ] |
|
|
