task stringclasses 9
values | state stringlengths 195 695 | instructions stringclasses 33
values | criteria_keys listlengths 2 5 | criteria_values listlengths 2 5 | label_alias stringclasses 5
values | label_key stringclasses 22
values | meta stringlengths 80 133 |
|---|---|---|---|---|---|---|---|
ticket.route | {"channel": "email", "ticket": {"subject": "Refund not received for cancelled sub", "body": "Thread summary — Invoice shows wrong amount on my card | wait_days=9 | tier=pro | priority=p2", "tier": "pro", "priority": "p2", "wait_days": 9}} | Route this support ticket to the best team. Reply with one option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"sales: pricing, upgrades, new accounts",
"billing: payment, charges, invoices, refunds",
"technical: bugs, crashes, outages, integrations",
"shipping: delivery, packages, tracking",
"other: unclear or general questions"
] | B | billing | {"gym": "ticket_dungeon", "domain_seed": "billing", "schema": "choice", "hard": false, "n_options": 5} |
ticket.route | {"channel": "phone", "ticket": {"subject": "Compare Business vs Enterprise plans", "body": "Thread summary — Can sales walk us through annual pricing? | wait_days=2 | tier=free | priority=p2", "tier": "free", "priority": "p2", "wait_days": 2}} | Choose the owning queue from Criteria. One letter only. | [
"A",
"B",
"C",
"D",
"E"
] | [
"other: unclear or general questions",
"sales: pricing, upgrades, new accounts",
"shipping: delivery, packages, tracking",
"technical: bugs, crashes, outages, integrations",
"billing: payment, charges, invoices, refunds"
] | B | sales | {"gym": "ticket_dungeon", "domain_seed": "sales", "schema": "choice", "hard": false, "n_options": 5} |
debate.confidence | {"question": "We should raise prices 10%", "side_b": {"label": "oppose", "claims": [{"text": "Churn spiked after last increase", "hard_evidence": true}, {"text": "Customers will be mad forever", "hard_evidence": false}]}, "side_a": {"label": "support", "claims": [{"text": "Gross margin is below target", "hard_evidence"... | Rate decision confidence from the evidence balance. One letter. | [
"A",
"B",
"C"
] | [
"0: low confidence",
"2: high confidence",
"1: medium confidence"
] | C | 1 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
alloc.fund_next | {"currency": "USD", "policy": "Fund at most one request per step; prefer higher urgency among affordable.", "budget_remaining": 223, "requests": [{"id": "compliance_audit", "title": "Compliance audit support", "cost": 67, "urgency": 3}, {"id": "design_sprint", "title": "Design sprint facilitator", "cost": 20, "urgency"... | Select the next funding target under the policy. One option letter. | [
"A",
"B",
"C",
"D"
] | [
"compliance_audit: Compliance audit support (cost=67, urgency=3)",
"hiring_bonus: Hiring referral bonuses (cost=59, urgency=1)",
"laptop_refresh: Engineer laptop refresh (cost=73, urgency=1)",
"design_sprint: Design sprint facilitator (cost=20, urgency=0)"
] | A | compliance_audit | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 4} |
alloc.pressure | {"requests": [{"id": "oncall_stipend", "title": "On-call stipend top-up", "cost": 43, "urgency": 2}, {"id": "hiring_bonus", "title": "Hiring referral bonuses", "cost": 55, "urgency": 1}, {"id": "compliance_audit", "title": "Compliance audit support", "cost": 77, "urgency": 3}, {"id": "security_patch", "title": "Critica... | How constrained is the budget relative to demand? Pick one letter. | [
"A",
"B"
] | [
"1: medium — close to capacity",
"2: high — demand exceeds budget"
] | A | 1 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 0.8087774294670846, "schema": "choice", "hard": false, "n_options": 2} |
alloc.pressure | {"currency": "USD", "policy": "Fund at most one request per step; prefer higher urgency among affordable.", "requests": [{"id": "laptop_refresh", "title": "Engineer laptop refresh", "cost": 72, "urgency": 1}, {"id": "marketing_campaign", "title": "Q3 marketing campaign", "cost": 110, "urgency": 1}, {"id": "security_pat... | Rate overall budget pressure. Reply with one level letter. | [
"A",
"B",
"C"
] | [
"1: medium — close to capacity",
"2: high — demand exceeds budget",
"0: low — budget comfortably covers demand"
] | B | 2 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 1.8412698412698412, "schema": "choice", "hard": false, "n_options": 3} |
ticket.needs_human | {"channel": "email", "ticket": {"subject": "Thanks — just saying hello", "body": "How do I change my display name? They have waited 2 days. Segment=pro.", "tier": "pro", "priority": "p0", "wait_days": 2}} | Should a human review this before automated handling? Reply with one letter. | [
"A",
"B"
] | [
"yes: Needs a human (money, legal, anger, long wait, outage)",
"no: Safe for bot / routine"
] | A | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
debate.confidence | {"side_b": {"label": "oppose", "claims": [{"text": "Enterprise deals are price-sensitive this quarter", "hard_evidence": true}, {"text": "Churn spiked after last increase", "hard_evidence": true}]}, "side_a": {"label": "support", "claims": [{"text": "Competitors already priced 12% higher", "hard_evidence": true}]}, "qu... | How confident should we be in the judgment? Reply with one level letter. | [
"A",
"B",
"C"
] | [
"1: medium confidence",
"2: high confidence",
"0: low confidence"
] | A | 1 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
ticket.urgency | {"channel": "chat", "ticket": {"subject": "Duplicate card charge on order A-590", "body": "Customer says: Invoice shows wrong amount on my card. Wait time so far: 7 days. Tier: free.", "tier": "free", "priority": "p0", "wait_days": 7}} | Rate urgency for this ticket on the given scale. Reply with one option letter. | [
"A",
"B",
"C"
] | [
"0: low — can wait",
"1: medium — same day",
"2: high — immediate"
] | C | 2 | {"gym": "ticket_dungeon", "schema_hint": "score", "urgency": 2, "schema": "choice", "hard": false, "n_options": 3} |
debate.enough_evidence | {"rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "side_b": {"label": "oppose", "claims": [{"text": "On-call coverage is thin on weekends", "hard_evidence": true}]}, "question": "We should deploy on Friday", "side_a": {"label": "support", "claims": [{"text": "Fridays fee... | Evidence sufficiency check — yes or no via option letter. | [
"A",
"B"
] | [
"yes: Enough hard evidence to decide",
"no: Too thin or tied — needs more evidence"
] | B | no | {"gym": "debate_judge", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
alloc.fund_next | {"currency": "USD", "policy": "Fund at most one request per step; prefer higher urgency among affordable.", "requests": [{"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 69, "urgency": 2}, {"id": "ml_experiment", "title": "GPU hours for ML experiment", "cost": 100, "urgency": 1}, {"id": "hirin... | Pick the best request to fund now. Prefer high urgency that fits the remaining budget. | [
"A",
"B",
"C"
] | [
"hiring_bonus: Hiring referral bonuses (cost=56, urgency=1)",
"customer_refund_pool: Customer refund reserve (cost=69, urgency=2)",
"ml_experiment: GPU hours for ML experiment (cost=100, urgency=1)"
] | B | customer_refund_pool | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 3} |
alloc.can_fund_all | {"budget_remaining": 78, "currency": "USD", "requests": [{"id": "security_patch", "title": "Critical security patch", "cost": 36, "urgency": 3}, {"id": "hiring_bonus", "title": "Hiring referral bonuses", "cost": 40, "urgency": 1}, {"id": "laptop_refresh", "title": "Engineer laptop refresh", "cost": 71, "urgency": 1}], ... | Would summing all costs fit under budget_remaining? Answer with one letter. | [
"A",
"B"
] | [
"yes: Budget covers the sum of all request costs",
"no: Budget is insufficient for all requests"
] | B | no | {"gym": "resource_allocator", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
ticket.route | {"channel": "phone", "ticket": {"subject": "Request a demo and pricing quote", "body": "Customer says: What plan is best for a 200 person team?. Wait time so far: 7 days. Tier: enterprise.", "tier": "enterprise", "priority": "p3", "wait_days": 7}} | Which team should own this ticket? Reply with exactly one option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"billing: payment, charges, invoices, refunds",
"sales: pricing, upgrades, new accounts",
"shipping: delivery, packages, tracking",
"other: unclear or general questions",
"technical: bugs, crashes, outages, integrations"
] | B | sales | {"gym": "ticket_dungeon", "domain_seed": "sales", "schema": "choice", "hard": false, "n_options": 5} |
ticket.urgency | {"ticket": {"subject": "General question about account", "body": "Thread summary — How do I change my display name? | wait_days=7 | tier=enterprise | priority=p1", "tier": "enterprise", "priority": "p1", "wait_days": 7}, "channel": "email"} | Assign an urgency score using the listed levels. One letter. | [
"A",
"B",
"C"
] | [
"1: medium — same day",
"2: high — immediate",
"0: low — can wait"
] | B | 2 | {"gym": "ticket_dungeon", "schema_hint": "score", "urgency": 2, "schema": "choice", "hard": false, "n_options": 3} |
ticket.needs_human | {"channel": "phone", "ticket": {"subject": "Interested in upgrading to Pro", "body": "Message: We need a quote for upgrading seats. Priority flag: p1. Account tier: enterprise.", "tier": "enterprise", "priority": "p1", "wait_days": 7}} | Does this ticket need a human agent (not a bot)? Reply with one option letter. | [
"A",
"B"
] | [
"no: Safe for bot / routine",
"yes: Needs a human (money, legal, anger, long wait, outage)"
] | B | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
alloc.fund_next | {"currency": "USD", "requests": [{"id": "hiring_bonus", "title": "Hiring referral bonuses", "cost": 42, "urgency": 1}, {"id": "oncall_stipend", "title": "On-call stipend top-up", "cost": 44, "urgency": 2}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 65, "urgency": 2}, {"id": "cdn_upgrade"... | Select the next funding target under the policy. One option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"customer_refund_pool: Customer refund reserve (cost=65, urgency=2)",
"oncall_stipend: On-call stipend top-up (cost=44, urgency=2)",
"hiring_bonus: Hiring referral bonuses (cost=42, urgency=1)",
"cdn_upgrade: CDN capacity upgrade (cost=51, urgency=2)",
"ml_experiment: GPU hours for ML experiment (cost=81, u... | B | oncall_stipend | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 5} |
debate.winner | {"side_a": {"label": "support", "claims": [{"text": "Open source is always good actually", "hard_evidence": false}, {"text": "Inbound PRs already fix bugs we lack time for", "hard_evidence": true}, {"text": "Design partners ask for source escrow", "hard_evidence": true}]}, "side_b": {"label": "oppose", "claims": [{"tex... | Apply the rubric: hard evidence beats vibes. Pick the winning side letter. | [
"A",
"B"
] | [
"oppose: Side B — oppose the proposal",
"support: Side A — support the proposal"
] | B | support | {"gym": "debate_judge", "pro_ev": 2, "con_ev": 2, "schema": "choice", "hard": false, "n_options": 2} |
alloc.pressure | {"requests": [{"id": "cdn_upgrade", "title": "CDN capacity upgrade", "cost": 56, "urgency": 2}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 54, "urgency": 2}, {"id": "hiring_bonus", "title": "Hiring referral bonuses", "cost": 41, "urgency": 1}, {"id": "compliance_audit", "title": "Complia... | Rate overall budget pressure. Reply with one level letter. | [
"A",
"B",
"C"
] | [
"2: high — demand exceeds budget",
"0: low — budget comfortably covers demand",
"1: medium — close to capacity"
] | C | 1 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 0.875, "schema": "choice", "hard": false, "n_options": 3} |
ticket.route | {"channel": "phone", "ticket": {"subject": "Request a demo and pricing quote", "body": "From customer (pro): Interested in enterprise pricing and a demo. SLA wait=3d priority=p1.", "tier": "pro", "priority": "p1", "wait_days": 3}} | Which team should own this ticket? Reply with exactly one option letter. | [
"A",
"B",
"C",
"D"
] | [
"technical: bugs, crashes, outages, integrations",
"sales: pricing, upgrades, new accounts",
"shipping: delivery, packages, tracking",
"other: unclear or general questions"
] | B | sales | {"gym": "ticket_dungeon", "domain_seed": "sales", "schema": "choice", "hard": false, "n_options": 4} |
ticket.route | {"ticket": {"subject": "Interested in upgrading to Pro", "body": "Interested in enterprise pricing and a demo They have waited 7 days. Segment=pro.", "tier": "pro", "priority": "p3", "wait_days": 7}, "channel": "chat"} | Choose the owning queue from Criteria. One letter only. | [
"A",
"B",
"C",
"D",
"E"
] | [
"sales: pricing, upgrades, new accounts",
"shipping: delivery, packages, tracking",
"technical: bugs, crashes, outages, integrations",
"billing: payment, charges, invoices, refunds",
"other: unclear or general questions"
] | A | sales | {"gym": "ticket_dungeon", "domain_seed": "sales", "schema": "choice", "hard": false, "n_options": 5} |
ticket.route | {"channel": "email", "ticket": {"subject": "Where is the documentation?", "body": "Customer says: Is there a status page?. Wait time so far: 6 days. Tier: enterprise.", "tier": "enterprise", "priority": "p0", "wait_days": 6}} | Route this support ticket to the best team. Reply with one option letter. | [
"A",
"B",
"C",
"D"
] | [
"shipping: delivery, packages, tracking",
"sales: pricing, upgrades, new accounts",
"other: unclear or general questions",
"technical: bugs, crashes, outages, integrations"
] | C | other | {"gym": "ticket_dungeon", "domain_seed": "other", "schema": "choice", "hard": false, "n_options": 4} |
debate.confidence | {"side_a": {"label": "support", "claims": [{"text": "Customers asked for the fix this week", "hard_evidence": true}, {"text": "Staging tests are green", "hard_evidence": true}]}, "question": "We should deploy on Friday", "rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "... | Rate decision confidence from the evidence balance. One letter. | [
"A",
"B",
"C"
] | [
"1: medium confidence",
"0: low confidence",
"2: high confidence"
] | A | 1 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
alloc.fund_next | {"requests": [{"id": "hiring_bonus", "title": "Hiring referral bonuses", "cost": 57, "urgency": 1}, {"id": "marketing_campaign", "title": "Q3 marketing campaign", "cost": 114, "urgency": 1}, {"id": "cdn_upgrade", "title": "CDN capacity upgrade", "cost": 49, "urgency": 2}, {"id": "office_snacks", "title": "Office snacks... | Select the next funding target under the policy. One option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"cdn_upgrade: CDN capacity upgrade (cost=49, urgency=2)",
"marketing_campaign: Q3 marketing campaign (cost=114, urgency=1)",
"office_snacks: Office snacks (cost=25, urgency=0)",
"hiring_bonus: Hiring referral bonuses (cost=57, urgency=1)",
"docs_rewrite: Docs rewrite contractor (cost=41, urgency=0)"
] | A | cdn_upgrade | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 5} |
alloc.pressure | {"budget_remaining": 137, "policy": "Fund at most one request per step; prefer higher urgency among affordable.", "currency": "USD", "requests": [{"id": "ml_experiment", "title": "GPU hours for ML experiment", "cost": 100, "urgency": 1}, {"id": "office_snacks", "title": "Office snacks", "cost": 10, "urgency": 0}, {"id"... | How constrained is the budget relative to demand? Pick one letter. | [
"A",
"B",
"C"
] | [
"1: medium — close to capacity",
"2: high — demand exceeds budget",
"0: low — budget comfortably covers demand"
] | A | 1 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 1.102189781021898, "schema": "choice", "hard": false, "n_options": 3} |
debate.confidence | {"rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "question": "We should open-source the core library", "side_a": {"label": "support", "claims": [{"text": "Design partners ask for source escrow", "hard_evidence": true}, {"text": "Open source is always good actually", "ha... | Rate decision confidence from the evidence balance. One letter. | [
"A",
"B"
] | [
"2: high confidence",
"0: low confidence"
] | B | 0 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 2} |
alloc.can_fund_all | {"requests": [{"id": "security_patch", "title": "Critical security patch", "cost": 31, "urgency": 3}, {"id": "ml_experiment", "title": "GPU hours for ML experiment", "cost": 89, "urgency": 1}, {"id": "marketing_campaign", "title": "Q3 marketing campaign", "cost": 129, "urgency": 1}], "currency": "USD", "policy": "Fund ... | Is the budget enough to approve every request? Reply with one option letter. | [
"A",
"B"
] | [
"no: Budget is insufficient for all requests",
"yes: Budget covers the sum of all request costs"
] | A | no | {"gym": "resource_allocator", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
debate.winner | {"question": "Remote-first improves productivity", "rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "side_a": {"label": "support", "claims": [{"text": "Hiring pipeline widened internationally", "hard_evidence": true}, {"text": "Attrition dropped 20% after remote policy",... | Given the claims and evidence flags, who should win the debate? One letter. | [
"A",
"B"
] | [
"support: Side A — support the proposal",
"oppose: Side B — oppose the proposal"
] | A | support | {"gym": "debate_judge", "pro_ev": 2, "con_ev": 2, "schema": "choice", "hard": false, "n_options": 2} |
alloc.pressure | {"requests": [{"id": "docs_rewrite", "title": "Docs rewrite contractor", "cost": 25, "urgency": 0}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 57, "urgency": 2}, {"id": "marketing_campaign", "title": "Q3 marketing campaign", "cost": 128, "urgency": 1}], "currency": "USD", "budget_remaini... | Budget tightness score — choose one Criteria letter. | [
"A",
"B",
"C"
] | [
"1: medium — close to capacity",
"2: high — demand exceeds budget",
"0: low — budget comfortably covers demand"
] | C | 0 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 0.7636363636363637, "schema": "choice", "hard": false, "n_options": 3} |
ticket.route | {"ticket": {"subject": "Invoice shows unexpected billing amount", "body": "Message: Invoice shows wrong amount on my card. Priority flag: p1. Account tier: free.", "tier": "free", "priority": "p1", "wait_days": 10}, "channel": "phone"} | Which team should own this ticket? Reply with exactly one option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"shipping: delivery, packages, tracking",
"other: unclear or general questions",
"billing: payment, charges, invoices, refunds",
"technical: bugs, crashes, outages, integrations",
"sales: pricing, upgrades, new accounts"
] | C | billing | {"gym": "ticket_dungeon", "domain_seed": "billing", "schema": "choice", "hard": false, "n_options": 5} |
ticket.route | {"ticket": {"subject": "Refund not received for cancelled sub", "body": "Message: Billing portal shows an incorrect charge. Priority flag: p1. Account tier: enterprise.", "tier": "enterprise", "priority": "p1", "wait_days": 3}, "channel": "chat"} | Route this support ticket to the best team. Reply with one option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"technical: bugs, crashes, outages, integrations",
"shipping: delivery, packages, tracking",
"sales: pricing, upgrades, new accounts",
"other: unclear or general questions",
"billing: payment, charges, invoices, refunds"
] | E | billing | {"gym": "ticket_dungeon", "domain_seed": "billing", "schema": "choice", "hard": false, "n_options": 5} |
debate.winner | {"side_a": {"label": "support", "claims": [{"text": "Prices going up sounds cool", "hard_evidence": false}, {"text": "Competitors already priced 12% higher", "hard_evidence": true}]}, "side_b": {"label": "oppose", "claims": [{"text": "Enterprise deals are price-sensitive this quarter", "hard_evidence": true}, {"text": ... | Apply the rubric: hard evidence beats vibes. Pick the winning side letter. | [
"A",
"B"
] | [
"support: Side A — support the proposal",
"oppose: Side B — oppose the proposal"
] | B | oppose | {"gym": "debate_judge", "pro_ev": 1, "con_ev": 2, "schema": "choice", "hard": false, "n_options": 2} |
alloc.can_fund_all | {"budget_remaining": 116, "currency": "USD", "policy": "Fund at most one request per step; prefer higher urgency among affordable.", "requests": [{"id": "office_snacks", "title": "Office snacks", "cost": 6, "urgency": 0}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 53, "urgency": 2}, {"id... | Is the budget enough to approve every request? Reply with one option letter. | [
"A",
"B"
] | [
"no: Budget is insufficient for all requests",
"yes: Budget covers the sum of all request costs"
] | B | yes | {"gym": "resource_allocator", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
ticket.route | {"ticket": {"subject": "API returns 500 on /v1/invoices", "body": "Message: Login fails with an internal error. Priority flag: p2. Account tier: enterprise.", "tier": "enterprise", "priority": "p2", "wait_days": 2}, "channel": "phone"} | Choose the owning queue from Criteria. One letter only. | [
"A",
"B",
"C",
"D",
"E"
] | [
"sales: pricing, upgrades, new accounts",
"other: unclear or general questions",
"shipping: delivery, packages, tracking",
"billing: payment, charges, invoices, refunds",
"technical: bugs, crashes, outages, integrations"
] | E | technical | {"gym": "ticket_dungeon", "domain_seed": "technical", "schema": "choice", "hard": false, "n_options": 5} |
alloc.pressure | {"requests": [{"id": "oncall_stipend", "title": "On-call stipend top-up", "cost": 46, "urgency": 2}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 69, "urgency": 2}, {"id": "laptop_refresh", "title": "Engineer laptop refresh", "cost": 74, "urgency": 1}], "currency": "USD", "policy": "Fund a... | Budget tightness score — choose one Criteria letter. | [
"A",
"B",
"C"
] | [
"1: medium — close to capacity",
"2: high — demand exceeds budget",
"0: low — budget comfortably covers demand"
] | A | 1 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 0.9043062200956937, "schema": "choice", "hard": false, "n_options": 3} |
alloc.fund_next | {"policy": "Fund at most one request per step; prefer higher urgency among affordable.", "currency": "USD", "budget_remaining": 245, "requests": [{"id": "compliance_audit", "title": "Compliance audit support", "cost": 64, "urgency": 3}, {"id": "ml_experiment", "title": "GPU hours for ML experiment", "cost": 88, "urgenc... | Which single request should be funded next given the budget? Reply with one letter. | [
"A",
"B",
"C",
"D"
] | [
"ml_experiment: GPU hours for ML experiment (cost=88, urgency=1)",
"compliance_audit: Compliance audit support (cost=64, urgency=3)",
"marketing_campaign: Q3 marketing campaign (cost=122, urgency=1)",
"laptop_refresh: Engineer laptop refresh (cost=78, urgency=1)"
] | B | compliance_audit | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 4} |
alloc.fund_next | {"budget_remaining": 262, "requests": [{"id": "marketing_campaign", "title": "Q3 marketing campaign", "cost": 130, "urgency": 1}, {"id": "compliance_audit", "title": "Compliance audit support", "cost": 77, "urgency": 3}, {"id": "design_sprint", "title": "Design sprint facilitator", "cost": 29, "urgency": 0}, {"id": "se... | Select the next funding target under the policy. One option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"office_snacks: Office snacks (cost=7, urgency=0)",
"design_sprint: Design sprint facilitator (cost=29, urgency=0)",
"compliance_audit: Compliance audit support (cost=77, urgency=3)",
"security_patch: Critical security patch (cost=48, urgency=3)",
"marketing_campaign: Q3 marketing campaign (cost=130, urgenc... | D | security_patch | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 5} |
ticket.route | {"channel": "email", "ticket": {"subject": "Production outage: checkout broken", "body": "Customer says: Login fails with an internal error. Wait time so far: 3 days. Tier: free.", "tier": "free", "priority": "p0", "wait_days": 3}} | Which team should own this ticket? Reply with exactly one option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"sales: pricing, upgrades, new accounts",
"billing: payment, charges, invoices, refunds",
"shipping: delivery, packages, tracking",
"technical: bugs, crashes, outages, integrations",
"other: unclear or general questions"
] | D | technical | {"gym": "ticket_dungeon", "domain_seed": "technical", "schema": "choice", "hard": false, "n_options": 5} |
debate.winner | {"question": "Remote-first improves productivity", "side_b": {"label": "oppose", "claims": [{"text": "New-grad ramp surveys show slower onboarding", "hard_evidence": true}, {"text": "Some customers require on-site workshops", "hard_evidence": true}, {"text": "Offices look empty and sad", "hard_evidence": false}]}, "rub... | Who wins under the stated evidence policy? One Criteria letter. | [
"A",
"B"
] | [
"oppose: Side B — oppose the proposal",
"support: Side A — support the proposal"
] | A | oppose | {"gym": "debate_judge", "pro_ev": 1, "con_ev": 2, "schema": "choice", "hard": false, "n_options": 2} |
ticket.urgency | {"ticket": {"subject": "Request a demo and pricing quote", "body": "From customer (pro): We need a quote for upgrading seats. SLA wait=3d priority=p1.", "tier": "pro", "priority": "p1", "wait_days": 3}, "channel": "phone"} | Rate urgency for this ticket on the given scale. Reply with one option letter. | [
"A",
"B",
"C"
] | [
"1: medium — same day",
"0: low — can wait",
"2: high — immediate"
] | C | 2 | {"gym": "ticket_dungeon", "schema_hint": "score", "urgency": 2, "schema": "choice", "hard": false, "n_options": 3} |
debate.enough_evidence | {"side_b": {"label": "oppose", "claims": [{"text": "Team has 1 Rust-experienced engineer", "hard_evidence": true}, {"text": "I dream in garbage collectors", "hard_evidence": false}, {"text": "Rewrite estimated at 2 quarters with feature freeze", "hard_evidence": true}]}, "side_a": {"label": "support", "claims": [{"text... | Do we have sufficient evidence-backed claims (not vibes)? Reply with one letter. | [
"A",
"B"
] | [
"no: Too thin or tied — needs more evidence",
"yes: Enough hard evidence to decide"
] | A | no | {"gym": "debate_judge", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
alloc.pressure | {"budget_remaining": 197, "currency": "USD", "policy": "Fund at most one request per step; prefer higher urgency among affordable.", "requests": [{"id": "office_snacks", "title": "Office snacks", "cost": 25, "urgency": 0}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 50, "urgency": 2}, {"i... | Rate overall budget pressure. Reply with one level letter. | [
"A",
"B",
"C"
] | [
"2: high — demand exceeds budget",
"0: low — budget comfortably covers demand",
"1: medium — close to capacity"
] | B | 0 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 0.6548223350253807, "schema": "choice", "hard": false, "n_options": 3} |
debate.winner | {"side_b": {"label": "oppose", "claims": [{"text": "Core contains proprietary scoring IP", "hard_evidence": true}]}, "rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "side_a": {"label": "support", "claims": [{"text": "Inbound PRs already fix bugs we lack time for", "hard... | Apply the rubric: hard evidence beats vibes. Pick the winning side letter. | [
"A",
"B"
] | [
"support: Side A — support the proposal",
"oppose: Side B — oppose the proposal"
] | A | support | {"gym": "debate_judge", "pro_ev": 2, "con_ev": 1, "schema": "choice", "hard": false, "n_options": 2} |
alloc.pressure | {"currency": "USD", "requests": [{"id": "security_patch", "title": "Critical security patch", "cost": 50, "urgency": 3}, {"id": "cdn_upgrade", "title": "CDN capacity upgrade", "cost": 52, "urgency": 2}, {"id": "ml_experiment", "title": "GPU hours for ML experiment", "cost": 86, "urgency": 1}], "policy": "Fund at most o... | How constrained is the budget relative to demand? Pick one letter. | [
"A",
"B",
"C"
] | [
"2: high — demand exceeds budget",
"1: medium — close to capacity",
"0: low — budget comfortably covers demand"
] | C | 0 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 0.7673469387755102, "schema": "choice", "hard": false, "n_options": 3} |
ticket.needs_human | {"ticket": {"subject": "How do I change my display name?", "body": "Message: Where can I find documentation?. Priority flag: p0. Account tier: enterprise.", "tier": "enterprise", "priority": "p0", "wait_days": 7}, "channel": "phone"} | Human escalation required? Pick yes or no via the option letter. | [
"A",
"B"
] | [
"yes: Needs a human (money, legal, anger, long wait, outage)",
"no: Safe for bot / routine"
] | A | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
ticket.urgency | {"ticket": {"subject": "Courier says package damaged", "body": "From customer (enterprise): Courier lost my shipment. SLA wait=10d priority=p0.", "tier": "enterprise", "priority": "p0", "wait_days": 10}, "channel": "phone"} | Urgency classification — choose exactly one Criteria letter. | [
"A",
"B",
"C"
] | [
"1: medium — same day",
"2: high — immediate",
"0: low — can wait"
] | B | 2 | {"gym": "ticket_dungeon", "schema_hint": "score", "urgency": 2, "schema": "choice", "hard": false, "n_options": 3} |
debate.enough_evidence | {"rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "side_b": {"label": "oppose", "claims": [{"text": "I just don't like Fridays", "hard_evidence": false}]}, "question": "We should deploy on Friday", "side_a": {"label": "support", "claims": [{"text": "Customers asked for t... | Evidence sufficiency check — yes or no via option letter. | [
"A",
"B"
] | [
"no: Too thin or tied — needs more evidence",
"yes: Enough hard evidence to decide"
] | A | no | {"gym": "debate_judge", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
ticket.route | {"ticket": {"subject": "How do I change my display name?", "body": "Is there a status page? They have waited 3 days. Segment=pro.", "tier": "pro", "priority": "p1", "wait_days": 3}, "channel": "phone"} | Best team for this case? Answer with a single option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"sales: pricing, upgrades, new accounts",
"billing: payment, charges, invoices, refunds",
"shipping: delivery, packages, tracking",
"other: unclear or general questions",
"technical: bugs, crashes, outages, integrations"
] | D | other | {"gym": "ticket_dungeon", "domain_seed": "other", "schema": "choice", "hard": false, "n_options": 5} |
ticket.needs_human | {"channel": "email", "ticket": {"subject": "Need enterprise pricing for 200 seats", "body": "From customer (enterprise): Can sales walk us through annual pricing?. SLA wait=1d priority=p1.", "tier": "enterprise", "priority": "p1", "wait_days": 1}} | Does this ticket need a human agent (not a bot)? Reply with one option letter. | [
"A",
"B"
] | [
"yes: Needs a human (money, legal, anger, long wait, outage)",
"no: Safe for bot / routine"
] | B | no | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
debate.confidence | {"side_a": {"label": "support", "claims": [{"text": "My friend likes remote", "hard_evidence": false}]}, "side_b": {"label": "oppose", "claims": [{"text": "Offices look empty and sad", "hard_evidence": false}, {"text": "New-grad ramp surveys show slower onboarding", "hard_evidence": true}]}, "question": "Remote-first i... | Rate decision confidence from the evidence balance. One letter. | [
"A",
"B",
"C"
] | [
"1: medium confidence",
"2: high confidence",
"0: low confidence"
] | C | 0 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
ticket.needs_human | {"ticket": {"subject": "Compare Business vs Enterprise plans", "body": "From customer (pro): What plan is best for a 200 person team?. SLA wait=8d priority=p1.", "tier": "pro", "priority": "p1", "wait_days": 8}, "channel": "email"} | Should a human review this before automated handling? Reply with one letter. | [
"A",
"B"
] | [
"no: Safe for bot / routine",
"yes: Needs a human (money, legal, anger, long wait, outage)"
] | B | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
debate.confidence | {"side_a": {"label": "support", "claims": [{"text": "Rust is trendy on HN", "hard_evidence": false}, {"text": "Rust prototype cut CPU 35% in bench", "hard_evidence": true}]}, "rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "question": "We should rewrite the backend in R... | How confident should we be in the judgment? Reply with one level letter. | [
"A",
"B",
"C"
] | [
"2: high confidence",
"0: low confidence",
"1: medium confidence"
] | C | 1 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
ticket.needs_human | {"ticket": {"subject": "Request a demo and pricing quote", "body": "Thread summary — Interested in enterprise pricing and a demo | wait_days=8 | tier=free | priority=p2", "tier": "free", "priority": "p2", "wait_days": 8}, "channel": "email"} | Should a human review this before automated handling? Reply with one letter. | [
"A",
"B"
] | [
"yes: Needs a human (money, legal, anger, long wait, outage)",
"no: Safe for bot / routine"
] | A | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
ticket.needs_human | {"channel": "phone", "ticket": {"subject": "App crashes when opening settings", "body": "Thread summary — Production API outage on invoices endpoint | wait_days=7 | tier=free | priority=p3", "tier": "free", "priority": "p3", "wait_days": 7}} | Human escalation required? Pick yes or no via the option letter. | [
"A",
"B"
] | [
"yes: Needs a human (money, legal, anger, long wait, outage)",
"no: Safe for bot / routine"
] | A | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
alloc.fund_next | {"policy": "Fund at most one request per step; prefer higher urgency among affordable.", "currency": "USD", "budget_remaining": 404, "requests": [{"id": "hiring_bonus", "title": "Hiring referral bonuses", "cost": 47, "urgency": 1}, {"id": "marketing_campaign", "title": "Q3 marketing campaign", "cost": 117, "urgency": 1... | Pick the best request to fund now. Prefer high urgency that fits the remaining budget. | [
"A",
"B",
"C",
"D",
"E"
] | [
"oncall_stipend: On-call stipend top-up (cost=51, urgency=2)",
"marketing_campaign: Q3 marketing campaign (cost=117, urgency=1)",
"docs_rewrite: Docs rewrite contractor (cost=43, urgency=0)",
"hiring_bonus: Hiring referral bonuses (cost=47, urgency=1)",
"ml_experiment: GPU hours for ML experiment (cost=89, ... | A | oncall_stipend | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 5} |
ticket.needs_human | {"ticket": {"subject": "Invoice shows unexpected billing amount", "body": "Customer says: Please reverse the duplicate payment. Wait time so far: 10 days. Tier: enterprise.", "tier": "enterprise", "priority": "p3", "wait_days": 10}, "channel": "chat"} | Does this ticket need a human agent (not a bot)? Reply with one option letter. | [
"A",
"B"
] | [
"no: Safe for bot / routine",
"yes: Needs a human (money, legal, anger, long wait, outage)"
] | B | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
debate.confidence | {"rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "question": "We should rewrite the backend in Rust", "side_b": {"label": "oppose", "claims": [{"text": "Team has 1 Rust-experienced engineer", "hard_evidence": true}, {"text": "Rewrite estimated at 2 quarters with feature... | Confidence score for this adjudication — pick one Criteria letter. | [
"A",
"B",
"C"
] | [
"1: medium confidence",
"2: high confidence",
"0: low confidence"
] | A | 1 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
alloc.can_fund_all | {"requests": [{"id": "docs_rewrite", "title": "Docs rewrite contractor", "cost": 33, "urgency": 0}, {"id": "design_sprint", "title": "Design sprint facilitator", "cost": 32, "urgency": 0}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 64, "urgency": 2}, {"id": "hiring_bonus", "title": "Hiri... | Would summing all costs fit under budget_remaining? Answer with one letter. | [
"A",
"B"
] | [
"no: Budget is insufficient for all requests",
"yes: Budget covers the sum of all request costs"
] | A | no | {"gym": "resource_allocator", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
debate.winner | {"question": "We should deploy on Friday", "side_a": {"label": "support", "claims": [{"text": "Customers asked for the fix this week", "hard_evidence": true}, {"text": "Fridays feel lucky", "hard_evidence": false}]}, "rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "side... | Which side is better supported by evidence? Reply with one option letter. | [
"A",
"B"
] | [
"oppose: Side B — oppose the proposal",
"support: Side A — support the proposal"
] | A | oppose | {"gym": "debate_judge", "pro_ev": 1, "con_ev": 1, "schema": "choice", "hard": false, "n_options": 2} |
alloc.fund_next | {"requests": [{"id": "marketing_campaign", "title": "Q3 marketing campaign", "cost": 110, "urgency": 1}, {"id": "oncall_stipend", "title": "On-call stipend top-up", "cost": 35, "urgency": 2}, {"id": "security_patch", "title": "Critical security patch", "cost": 34, "urgency": 3}, {"id": "ml_experiment", "title": "GPU ho... | Which single request should be funded next given the budget? Reply with one letter. | [
"A",
"B",
"C",
"D"
] | [
"marketing_campaign: Q3 marketing campaign (cost=110, urgency=1)",
"security_patch: Critical security patch (cost=34, urgency=3)",
"oncall_stipend: On-call stipend top-up (cost=35, urgency=2)",
"ml_experiment: GPU hours for ML experiment (cost=100, urgency=1)"
] | B | security_patch | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 4} |
alloc.fund_next | {"policy": "Fund at most one request per step; prefer higher urgency among affordable.", "currency": "USD", "budget_remaining": 135, "requests": [{"id": "hiring_bonus", "title": "Hiring referral bonuses", "cost": 60, "urgency": 1}, {"id": "security_patch", "title": "Critical security patch", "cost": 36, "urgency": 3}, ... | Select the next funding target under the policy. One option letter. | [
"A",
"B",
"C",
"D"
] | [
"office_snacks: Office snacks (cost=11, urgency=0)",
"hiring_bonus: Hiring referral bonuses (cost=60, urgency=1)",
"security_patch: Critical security patch (cost=36, urgency=3)",
"design_sprint: Design sprint facilitator (cost=33, urgency=0)"
] | C | security_patch | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 4} |
ticket.urgency | {"channel": "email", "ticket": {"subject": "Need enterprise pricing for 200 seats", "body": "Thread summary — Can sales walk us through annual pricing? | wait_days=7 | tier=pro | priority=p0", "tier": "pro", "priority": "p0", "wait_days": 7}} | Assign an urgency score using the listed levels. One letter. | [
"A",
"B",
"C"
] | [
"2: high — immediate",
"1: medium — same day",
"0: low — can wait"
] | A | 2 | {"gym": "ticket_dungeon", "schema_hint": "score", "urgency": 2, "schema": "choice", "hard": false, "n_options": 3} |
debate.confidence | {"rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "side_b": {"label": "oppose", "claims": [{"text": "On-call coverage is thin on weekends", "hard_evidence": true}, {"text": "Last Friday deploy caused a Sev-1", "hard_evidence": true}]}, "side_a": {"label": "support", "cla... | Confidence score for this adjudication — pick one Criteria letter. | [
"A",
"B",
"C"
] | [
"2: high confidence",
"0: low confidence",
"1: medium confidence"
] | C | 1 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
alloc.pressure | {"currency": "USD", "requests": [{"id": "design_sprint", "title": "Design sprint facilitator", "cost": 33, "urgency": 0}, {"id": "oncall_stipend", "title": "On-call stipend top-up", "cost": 50, "urgency": 2}, {"id": "hiring_bonus", "title": "Hiring referral bonuses", "cost": 56, "urgency": 1}, {"id": "security_patch", ... | Rate overall budget pressure. Reply with one level letter. | [
"A",
"B",
"C"
] | [
"2: high — demand exceeds budget",
"1: medium — close to capacity",
"0: low — budget comfortably covers demand"
] | A | 2 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 1.3065693430656935, "schema": "choice", "hard": false, "n_options": 3} |
ticket.route | {"channel": "phone", "ticket": {"subject": "Invoice shows unexpected billing amount", "body": "From customer (enterprise): I was charged twice and want a refund immediately. SLA wait=0d priority=p0.", "tier": "enterprise", "priority": "p0", "wait_days": 0}} | Route this support ticket to the best team. Reply with one option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"technical: bugs, crashes, outages, integrations",
"billing: payment, charges, invoices, refunds",
"sales: pricing, upgrades, new accounts",
"shipping: delivery, packages, tracking",
"other: unclear or general questions"
] | B | billing | {"gym": "ticket_dungeon", "domain_seed": "billing", "schema": "choice", "hard": false, "n_options": 5} |
debate.confidence | {"side_a": {"label": "support", "claims": [{"text": "p99 latency is 4x SLO on hot path", "hard_evidence": true}]}, "question": "We should rewrite the backend in Rust", "rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "side_b": {"label": "oppose", "claims": [{"text": "I d... | Rate decision confidence from the evidence balance. One letter. | [
"A",
"B",
"C"
] | [
"0: low confidence",
"1: medium confidence",
"2: high confidence"
] | B | 1 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
ticket.needs_human | {"ticket": {"subject": "Package stuck in transit for 5 days", "body": "Thread summary — Courier lost my shipment | wait_days=5 | tier=pro | priority=p3", "tier": "pro", "priority": "p3", "wait_days": 5}, "channel": "email"} | Human escalation required? Pick yes or no via the option letter. | [
"A",
"B"
] | [
"no: Safe for bot / routine",
"yes: Needs a human (money, legal, anger, long wait, outage)"
] | B | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
ticket.urgency | {"channel": "phone", "ticket": {"subject": "Courier says package damaged", "body": "Package marked delivered but nothing arrived They have waited 5 days. Segment=free.", "tier": "free", "priority": "p0", "wait_days": 5}} | Rate urgency for this ticket on the given scale. Reply with one option letter. | [
"A",
"B"
] | [
"2: high — immediate",
"0: low — can wait"
] | A | 2 | {"gym": "ticket_dungeon", "schema_hint": "score", "urgency": 2, "schema": "choice", "hard": false, "n_options": 2} |
ticket.urgency | {"channel": "phone", "ticket": {"subject": "How do I change my display name?", "body": "Customer says: Thanks for the help last week. Wait time so far: 9 days. Tier: enterprise.", "tier": "enterprise", "priority": "p0", "wait_days": 9}} | Rate urgency for this ticket on the given scale. Reply with one option letter. | [
"A",
"B",
"C"
] | [
"2: high — immediate",
"1: medium — same day",
"0: low — can wait"
] | A | 2 | {"gym": "ticket_dungeon", "schema_hint": "score", "urgency": 2, "schema": "choice", "hard": false, "n_options": 3} |
alloc.fund_next | {"requests": [{"id": "docs_rewrite", "title": "Docs rewrite contractor", "cost": 40, "urgency": 0}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 58, "urgency": 2}, {"id": "office_snacks", "title": "Office snacks", "cost": 21, "urgency": 0}], "policy": "Fund at most one request per step; pr... | Which single request should be funded next given the budget? Reply with one letter. | [
"A",
"B",
"C"
] | [
"docs_rewrite: Docs rewrite contractor (cost=40, urgency=0)",
"office_snacks: Office snacks (cost=21, urgency=0)",
"customer_refund_pool: Customer refund reserve (cost=58, urgency=2)"
] | C | customer_refund_pool | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 3} |
alloc.fund_next | {"currency": "USD", "budget_remaining": 209, "policy": "Fund at most one request per step; prefer higher urgency among affordable.", "requests": [{"id": "security_patch", "title": "Critical security patch", "cost": 33, "urgency": 3}, {"id": "laptop_refresh", "title": "Engineer laptop refresh", "cost": 79, "urgency": 1}... | Given budget_remaining and urgencies, which request wins? One letter. | [
"A",
"B",
"C"
] | [
"oncall_stipend: On-call stipend top-up (cost=39, urgency=2)",
"laptop_refresh: Engineer laptop refresh (cost=79, urgency=1)",
"security_patch: Critical security patch (cost=33, urgency=3)"
] | C | security_patch | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 3} |
debate.confidence | {"rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "side_b": {"label": "oppose", "claims": [{"text": "I just don't like Fridays", "hard_evidence": false}, {"text": "On-call coverage is thin on weekends", "hard_evidence": true}]}, "question": "We should deploy on Friday", ... | How confident should we be in the judgment? Reply with one level letter. | [
"A",
"B",
"C"
] | [
"2: high confidence",
"1: medium confidence",
"0: low confidence"
] | C | 0 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
ticket.urgency | {"ticket": {"subject": "Request a demo and pricing quote", "body": "Message: What plan is best for a 200 person team?. Priority flag: p2. Account tier: free.", "tier": "free", "priority": "p2", "wait_days": 10}, "channel": "email"} | Urgency classification — choose exactly one Criteria letter. | [
"A",
"B",
"C"
] | [
"1: medium — same day",
"0: low — can wait",
"2: high — immediate"
] | C | 2 | {"gym": "ticket_dungeon", "schema_hint": "score", "urgency": 2, "schema": "choice", "hard": false, "n_options": 3} |
alloc.can_fund_all | {"currency": "USD", "requests": [{"id": "cdn_upgrade", "title": "CDN capacity upgrade", "cost": 56, "urgency": 2}, {"id": "laptop_refresh", "title": "Engineer laptop refresh", "cost": 84, "urgency": 1}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 62, "urgency": 2}, {"id": "design_sprint",... | Would summing all costs fit under budget_remaining? Answer with one letter. | [
"A",
"B"
] | [
"yes: Budget covers the sum of all request costs",
"no: Budget is insufficient for all requests"
] | A | yes | {"gym": "resource_allocator", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
ticket.needs_human | {"ticket": {"subject": "Courier says package damaged", "body": "Customer says: Tracking has not moved for a week. Wait time so far: 5 days. Tier: free.", "tier": "free", "priority": "p3", "wait_days": 5}, "channel": "phone"} | Human escalation required? Pick yes or no via the option letter. | [
"A",
"B"
] | [
"no: Safe for bot / routine",
"yes: Needs a human (money, legal, anger, long wait, outage)"
] | B | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
debate.confidence | {"rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "question": "We should hire more junior engineers", "side_b": {"label": "oppose", "claims": [{"text": "Critical path needs senior ownership now", "hard_evidence": true}]}, "side_a": {"label": "support", "claims": [{"text"... | How confident should we be in the judgment? Reply with one level letter. | [
"A",
"B",
"C"
] | [
"2: high confidence",
"0: low confidence",
"1: medium confidence"
] | B | 0 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
ticket.route | {"channel": "phone", "ticket": {"subject": "Courier says package damaged", "body": "Message: Package marked delivered but nothing arrived. Priority flag: p0. Account tier: enterprise.", "tier": "enterprise", "priority": "p0", "wait_days": 0}} | Choose the owning queue from Criteria. One letter only. | [
"A",
"B",
"C",
"D",
"E"
] | [
"billing: payment, charges, invoices, refunds",
"technical: bugs, crashes, outages, integrations",
"sales: pricing, upgrades, new accounts",
"other: unclear or general questions",
"shipping: delivery, packages, tracking"
] | E | shipping | {"gym": "ticket_dungeon", "domain_seed": "shipping", "schema": "choice", "hard": false, "n_options": 5} |
debate.confidence | {"rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "side_b": {"label": "oppose", "claims": [{"text": "Critical path needs senior ownership now", "hard_evidence": true}, {"text": "Juniors ask too many questions", "hard_evidence": false}]}, "side_a": {"label": "support", "c... | Confidence score for this adjudication — pick one Criteria letter. | [
"A",
"B",
"C"
] | [
"1: medium confidence",
"0: low confidence",
"2: high confidence"
] | B | 0 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
alloc.fund_next | {"currency": "USD", "budget_remaining": 386, "requests": [{"id": "ml_experiment", "title": "GPU hours for ML experiment", "cost": 90, "urgency": 1}, {"id": "docs_rewrite", "title": "Docs rewrite contractor", "cost": 26, "urgency": 0}, {"id": "security_patch", "title": "Critical security patch", "cost": 45, "urgency": 3... | Which single request should be funded next given the budget? Reply with one letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"security_patch: Critical security patch (cost=45, urgency=3)",
"marketing_campaign: Q3 marketing campaign (cost=112, urgency=1)",
"ml_experiment: GPU hours for ML experiment (cost=90, urgency=1)",
"customer_refund_pool: Customer refund reserve (cost=67, urgency=2)",
"docs_rewrite: Docs rewrite contractor (... | A | security_patch | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 5} |
alloc.can_fund_all | {"policy": "Fund at most one request per step; prefer higher urgency among affordable.", "budget_remaining": 251, "requests": [{"id": "security_patch", "title": "Critical security patch", "cost": 35, "urgency": 3}, {"id": "laptop_refresh", "title": "Engineer laptop refresh", "cost": 79, "urgency": 1}, {"id": "ml_experi... | Would summing all costs fit under budget_remaining? Answer with one letter. | [
"A",
"B"
] | [
"no: Budget is insufficient for all requests",
"yes: Budget covers the sum of all request costs"
] | A | no | {"gym": "resource_allocator", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
debate.enough_evidence | {"side_b": {"label": "oppose", "claims": [{"text": "Ramp cost exceeds budget this half", "hard_evidence": true}, {"text": "Critical path needs senior ownership now", "hard_evidence": true}]}, "side_a": {"label": "support", "claims": [{"text": "Senior mentors have spare capacity", "hard_evidence": true}, {"text": "Inter... | Evidence sufficiency check — yes or no via option letter. | [
"A",
"B"
] | [
"no: Too thin or tied — needs more evidence",
"yes: Enough hard evidence to decide"
] | A | no | {"gym": "debate_judge", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
ticket.needs_human | {"channel": "chat", "ticket": {"subject": "Invoice shows unexpected billing amount", "body": "Please reverse the duplicate payment They have waited 9 days. Segment=free.", "tier": "free", "priority": "p2", "wait_days": 9}} | Should a human review this before automated handling? Reply with one letter. | [
"A",
"B"
] | [
"no: Safe for bot / routine",
"yes: Needs a human (money, legal, anger, long wait, outage)"
] | B | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
ticket.route | {"ticket": {"subject": "Charged twice for the same invoice A-177", "body": "Thread summary — Billing portal shows an incorrect charge | wait_days=7 | tier=pro | priority=p0", "tier": "pro", "priority": "p0", "wait_days": 7}, "channel": "phone"} | Route this support ticket to the best team. Reply with one option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"billing: payment, charges, invoices, refunds",
"other: unclear or general questions",
"technical: bugs, crashes, outages, integrations",
"shipping: delivery, packages, tracking",
"sales: pricing, upgrades, new accounts"
] | A | billing | {"gym": "ticket_dungeon", "domain_seed": "billing", "schema": "choice", "hard": false, "n_options": 5} |
alloc.pressure | {"policy": "Fund at most one request per step; prefer higher urgency among affordable.", "budget_remaining": 83, "currency": "USD", "requests": [{"id": "docs_rewrite", "title": "Docs rewrite contractor", "cost": 39, "urgency": 0}, {"id": "laptop_refresh", "title": "Engineer laptop refresh", "cost": 77, "urgency": 1}, {... | How constrained is the budget relative to demand? Pick one letter. | [
"A",
"B",
"C"
] | [
"0: low — budget comfortably covers demand",
"2: high — demand exceeds budget",
"1: medium — close to capacity"
] | B | 2 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 1.819277108433735, "schema": "choice", "hard": false, "n_options": 3} |
alloc.fund_next | {"policy": "Fund at most one request per step; prefer higher urgency among affordable.", "requests": [{"id": "oncall_stipend", "title": "On-call stipend top-up", "cost": 36, "urgency": 2}, {"id": "design_sprint", "title": "Design sprint facilitator", "cost": 17, "urgency": 0}, {"id": "marketing_campaign", "title": "Q3 ... | Given budget_remaining and urgencies, which request wins? One letter. | [
"A",
"B",
"C",
"D"
] | [
"office_snacks: Office snacks (cost=7, urgency=0)",
"design_sprint: Design sprint facilitator (cost=17, urgency=0)",
"oncall_stipend: On-call stipend top-up (cost=36, urgency=2)",
"marketing_campaign: Q3 marketing campaign (cost=127, urgency=1)"
] | C | oncall_stipend | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 4} |
ticket.needs_human | {"channel": "phone", "ticket": {"subject": "Delivery marked complete but missing", "body": "Courier lost my shipment They have waited 8 days. Segment=enterprise.", "tier": "enterprise", "priority": "p0", "wait_days": 8}} | Human escalation required? Pick yes or no via the option letter. | [
"A",
"B"
] | [
"yes: Needs a human (money, legal, anger, long wait, outage)",
"no: Safe for bot / routine"
] | A | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
alloc.fund_next | {"currency": "USD", "budget_remaining": 100, "requests": [{"id": "cdn_upgrade", "title": "CDN capacity upgrade", "cost": 49, "urgency": 2}, {"id": "security_patch", "title": "Critical security patch", "cost": 46, "urgency": 3}, {"id": "docs_rewrite", "title": "Docs rewrite contractor", "cost": 27, "urgency": 0}, {"id":... | Pick the best request to fund now. Prefer high urgency that fits the remaining budget. | [
"A",
"B",
"C",
"D"
] | [
"cdn_upgrade: CDN capacity upgrade (cost=49, urgency=2)",
"docs_rewrite: Docs rewrite contractor (cost=27, urgency=0)",
"security_patch: Critical security patch (cost=46, urgency=3)",
"customer_refund_pool: Customer refund reserve (cost=67, urgency=2)"
] | C | security_patch | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 4} |
alloc.can_fund_all | {"policy": "Fund at most one request per step; prefer higher urgency among affordable.", "currency": "USD", "budget_remaining": 203, "requests": [{"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 62, "urgency": 2}, {"id": "design_sprint", "title": "Design sprint facilitator", "cost": 19, "urgen... | Is the budget enough to approve every request? Reply with one option letter. | [
"A",
"B"
] | [
"yes: Budget covers the sum of all request costs",
"no: Budget is insufficient for all requests"
] | B | no | {"gym": "resource_allocator", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
ticket.needs_human | {"ticket": {"subject": "General question about account", "body": "Customer says: Is there a status page?. Wait time so far: 9 days. Tier: enterprise.", "tier": "enterprise", "priority": "p2", "wait_days": 9}, "channel": "email"} | Should a human review this before automated handling? Reply with one letter. | [
"A",
"B"
] | [
"yes: Needs a human (money, legal, anger, long wait, outage)",
"no: Safe for bot / routine"
] | A | yes | {"gym": "ticket_dungeon", "schema_hint": "noul", "schema": "choice", "hard": false, "n_options": 2} |
debate.confidence | {"side_a": {"label": "support", "claims": [{"text": "Internship NPS was high and conversion strong", "hard_evidence": true}, {"text": "Senior mentors have spare capacity", "hard_evidence": true}, {"text": "Juniors are cheaper vibes", "hard_evidence": false}]}, "rubric": "Prefer hard evidence over vibes; break ties towa... | Confidence score for this adjudication — pick one Criteria letter. | [
"A",
"B",
"C"
] | [
"2: high confidence",
"0: low confidence",
"1: medium confidence"
] | B | 0 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
ticket.route | {"ticket": {"subject": "Need enterprise pricing for 200 seats", "body": "Thread summary — What plan is best for a 200 person team? | wait_days=0 | tier=pro | priority=p2", "tier": "pro", "priority": "p2", "wait_days": 0}, "channel": "phone"} | Which team should own this ticket? Reply with exactly one option letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"shipping: delivery, packages, tracking",
"sales: pricing, upgrades, new accounts",
"technical: bugs, crashes, outages, integrations",
"other: unclear or general questions",
"billing: payment, charges, invoices, refunds"
] | B | sales | {"gym": "ticket_dungeon", "domain_seed": "sales", "schema": "choice", "hard": false, "n_options": 5} |
debate.confidence | {"side_b": {"label": "oppose", "claims": [{"text": "Support load estimates exceed current headcount", "hard_evidence": true}]}, "question": "We should open-source the core library", "side_a": {"label": "support", "claims": [{"text": "Inbound PRs already fix bugs we lack time for", "hard_evidence": true}]}, "rubric": "P... | Rate decision confidence from the evidence balance. One letter. | [
"A",
"B",
"C"
] | [
"0: low confidence",
"2: high confidence",
"1: medium confidence"
] | A | 0 | {"gym": "debate_judge", "schema_hint": "score", "schema": "choice", "hard": false, "n_options": 3} |
alloc.fund_next | {"policy": "Fund at most one request per step; prefer higher urgency among affordable.", "budget_remaining": 185, "currency": "USD", "requests": [{"id": "laptop_refresh", "title": "Engineer laptop refresh", "cost": 82, "urgency": 1}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 66, "urgenc... | Pick the best request to fund now. Prefer high urgency that fits the remaining budget. | [
"A",
"B",
"C"
] | [
"compliance_audit: Compliance audit support (cost=73, urgency=3)",
"customer_refund_pool: Customer refund reserve (cost=66, urgency=2)",
"laptop_refresh: Engineer laptop refresh (cost=82, urgency=1)"
] | A | compliance_audit | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 3} |
alloc.pressure | {"budget_remaining": 414, "requests": [{"id": "cdn_upgrade", "title": "CDN capacity upgrade", "cost": 53, "urgency": 2}, {"id": "ml_experiment", "title": "GPU hours for ML experiment", "cost": 92, "urgency": 1}, {"id": "compliance_audit", "title": "Compliance audit support", "cost": 65, "urgency": 3}, {"id": "marketing... | How constrained is the budget relative to demand? Pick one letter. | [
"A",
"B",
"C"
] | [
"1: medium — close to capacity",
"0: low — budget comfortably covers demand",
"2: high — demand exceeds budget"
] | A | 1 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 0.9009661835748792, "schema": "choice", "hard": false, "n_options": 3} |
alloc.fund_next | {"requests": [{"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 60, "urgency": 2}, {"id": "compliance_audit", "title": "Compliance audit support", "cost": 60, "urgency": 3}, {"id": "design_sprint", "title": "Design sprint facilitator", "cost": 24, "urgency": 0}, {"id": "ml_experiment", "title":... | Pick the best request to fund now. Prefer high urgency that fits the remaining budget. | [
"A",
"B",
"C",
"D"
] | [
"ml_experiment: GPU hours for ML experiment (cost=93, urgency=1)",
"customer_refund_pool: Customer refund reserve (cost=60, urgency=2)",
"compliance_audit: Compliance audit support (cost=60, urgency=3)",
"design_sprint: Design sprint facilitator (cost=24, urgency=0)"
] | C | compliance_audit | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 4} |
alloc.pressure | {"budget_remaining": 393, "policy": "Fund at most one request per step; prefer higher urgency among affordable.", "requests": [{"id": "docs_rewrite", "title": "Docs rewrite contractor", "cost": 44, "urgency": 0}, {"id": "marketing_campaign", "title": "Q3 marketing campaign", "cost": 111, "urgency": 1}, {"id": "hiring_b... | How constrained is the budget relative to demand? Pick one letter. | [
"A",
"B",
"C"
] | [
"2: high — demand exceeds budget",
"1: medium — close to capacity",
"0: low — budget comfortably covers demand"
] | B | 1 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 0.8575063613231552, "schema": "choice", "hard": false, "n_options": 3} |
debate.winner | {"question": "We should hire more junior engineers", "rubric": "Prefer hard evidence over vibes; break ties toward the side with more hard evidence.", "side_a": {"label": "support", "claims": [{"text": "Senior mentors have spare capacity", "hard_evidence": true}, {"text": "Internship NPS was high and conversion strong"... | Given the claims and evidence flags, who should win the debate? One letter. | [
"A",
"B"
] | [
"support: Side A — support the proposal",
"oppose: Side B — oppose the proposal"
] | B | oppose | {"gym": "debate_judge", "pro_ev": 2, "con_ev": 2, "schema": "choice", "hard": false, "n_options": 2} |
ticket.urgency | {"ticket": {"subject": "App crashes when opening settings", "body": "From customer (enterprise): Stack trace on checkout — looks like a bug. SLA wait=9d priority=p2.", "tier": "enterprise", "priority": "p2", "wait_days": 9}, "channel": "phone"} | Assign an urgency score using the listed levels. One letter. | [
"A",
"B",
"C"
] | [
"1: medium — same day",
"2: high — immediate",
"0: low — can wait"
] | B | 2 | {"gym": "ticket_dungeon", "schema_hint": "score", "urgency": 2, "schema": "choice", "hard": false, "n_options": 3} |
alloc.fund_next | {"currency": "USD", "policy": "Fund at most one request per step; prefer higher urgency among affordable.", "requests": [{"id": "oncall_stipend", "title": "On-call stipend top-up", "cost": 54, "urgency": 2}, {"id": "customer_refund_pool", "title": "Customer refund reserve", "cost": 62, "urgency": 2}, {"id": "cdn_upgrad... | Given budget_remaining and urgencies, which request wins? One letter. | [
"A",
"B",
"C",
"D",
"E"
] | [
"hiring_bonus: Hiring referral bonuses (cost=54, urgency=1)",
"oncall_stipend: On-call stipend top-up (cost=54, urgency=2)",
"cdn_upgrade: CDN capacity upgrade (cost=56, urgency=2)",
"customer_refund_pool: Customer refund reserve (cost=62, urgency=2)",
"ml_experiment: GPU hours for ML experiment (cost=83, u... | B | oncall_stipend | {"gym": "resource_allocator", "schema": "choice", "hard": false, "n_options": 5} |
ticket.urgency | {"channel": "phone", "ticket": {"subject": "Need enterprise pricing for 200 seats", "body": "Message: Can sales walk us through annual pricing?. Priority flag: p1. Account tier: pro.", "tier": "pro", "priority": "p1", "wait_days": 2}} | Rate urgency for this ticket on the given scale. Reply with one option letter. | [
"A",
"B",
"C"
] | [
"1: medium — same day",
"0: low — can wait",
"2: high — immediate"
] | A | 1 | {"gym": "ticket_dungeon", "schema_hint": "score", "urgency": 1, "schema": "choice", "hard": false, "n_options": 3} |
alloc.pressure | {"requests": [{"id": "office_snacks", "title": "Office snacks", "cost": 9, "urgency": 0}, {"id": "hiring_bonus", "title": "Hiring referral bonuses", "cost": 44, "urgency": 1}, {"id": "cdn_upgrade", "title": "CDN capacity upgrade", "cost": 52, "urgency": 2}, {"id": "compliance_audit", "title": "Compliance audit support"... | Rate overall budget pressure. Reply with one level letter. | [
"A",
"B",
"C"
] | [
"1: medium — close to capacity",
"0: low — budget comfortably covers demand",
"2: high — demand exceeds budget"
] | B | 0 | {"gym": "resource_allocator", "schema_hint": "score", "ratio": 0.6988416988416989, "schema": "choice", "hard": false, "n_options": 3} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.