# s1mb-generalization-contextual-choice

English evaluation cases authored using GPT-6-Astra. This review contains answers; do not send it as model input.

## contextual-choice-001

### Instruction

```json
"What should we serve this guest, using only their stated requirements?"
```

### State

```json
[
  "The guest is fasting until sunset.",
  "It is noon; they explicitly decline food until the fast ends."
]
```

### Criteria

- `option_1` (value: None): "Cedar-certified chicken plate"
- `option_2` (value: None): "No meal until sunset"
- `option_3` (value: None): "Fish with rice"
- `option_4` (value: None): "Lentil stew with vegetable stock"
- `option_5` (value: None): "Plain rice and sealed water"
- `option_6` (value: None): "Cheese sandwich"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0}

Deferring the meal respects the stated fast.

## contextual-choice-002

### Instruction

```json
"Should the application enable this sharing feature now?"
```

### State

```json
"The policy requires affirmative consent. The user skipped the consent screen; defaults do not count as consent."
```

### Criteria

- `option_1` (value: None): "Enable sharing"
- `option_2` (value: None): "Keep sharing disabled"

Target: {'option_1': 0.0, 'option_2': 1.0}

Skipping supplies no affirmative consent.

## contextual-choice-003

### Instruction

```json
"Should the application enable this sharing feature now?"
```

### State

```json
{
  "policy": "Explicit opt-in required.",
  "user_action": "Checked the opt-in box and saved."
}
```

### Criteria

- `option_1` (value: None): "Enable sharing"
- `option_2` (value: None): "Keep sharing disabled"

Target: {'option_1': 1.0, 'option_2': 0.0}

A saved explicit opt-in grants permission.

## contextual-choice-004

### Instruction

```json
"Should the application enable this sharing feature now?"
```

### State

```json
{
  "permission": "Active until tomorrow; not revoked.",
  "policy": "Sharing enabled while a user-set permission is active."
}
```

### Criteria

- `option_1` (value: None): "Enable sharing"
- `option_2` (value: None): "Keep sharing disabled"

Target: {'option_1': 1.0, 'option_2': 0.0}

The permission remains active now.

## contextual-choice-005

### Instruction

```json
"Which event format respects the organizer's stated commitments?"
```

### State

```json
"The organizer's temporary vow requires solitude at home and no online participation today. They still want personal reflection."
```

### Criteria

- `option_1` (value: None): "Interfaith dialogue circle"
- `option_2` (value: None): "Silent observation of a rite"
- `option_3` (value: None): "Postpone the event"
- `option_4` (value: None): "Outdoor communal singing"
- `option_5` (value: None): "Participatory prayer"
- `option_6` (value: None): "Secular history lecture"
- `option_7` (value: None): "Private reflection at home"
- `option_8` (value: None): "Online text discussion"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 1.0, 'option_8': 0.0}

Private reflection at home honors solitude and location.

## contextual-choice-006

### Instruction

```json
"How should the facilitator resolve this decision under the group's charter?"
```

### State

```json
{
  "charter": "Constitutional amendments require every member's consent.",
  "decision": "A constitutional amendment; quorum present."
}
```

### Criteria

- `option_1` (value: None): "Require unanimous consent"
- `option_2` (value: None): "Defer to the affected subgroup"
- `option_3` (value: None): "Let the rotating steward decide"
- `option_4` (value: None): "Draw lots"
- `option_5` (value: None): "Simple majority vote"
- `option_6` (value: None): "Seek external arbitration"
- `option_7` (value: None): "Run a reversible pilot"
- `option_8` (value: None): "Postpone until quorum"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0, 'option_8': 0.0}

Universal consent is required for this class of decision.

## contextual-choice-007

### Instruction

```json
"Which version best matches the requested tone and meaning?"
```

### State

```json
{
  "intent": "State permission to send today, not a request or obligation."
}
```

### Criteria

- `option_1` (value: None): "Would you mind sending the report today?"
- `option_2` (value: None): "The report may be sent today."
- `option_3` (value: None): "Do not send the report today."
- `option_4` (value: None): "Send the report today."
- `option_5` (value: None): "Please send the report today."

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

May expresses permission.

## contextual-choice-008

### Instruction

```json
"What is the appropriate release format under this agreement?"
```

### State

```json
{
  "agreement": "Release only group counts, with no row-level records.",
  "request": "Approved statistical report; all group sizes meet the agreed minimum."
}
```

### Criteria

- `option_1` (value: None): "Identifiable rows"
- `option_2` (value: None): "Pseudonymized rows"
- `option_3` (value: None): "Synthetic sample"
- `option_4` (value: None): "Do not release"
- `option_5` (value: None): "Aggregated counts"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0}

Aggregated counts meet the row-level restriction.

## contextual-choice-009

### Instruction

```json
"Which bid should procurement select under the stated rule?"
```

### State

```json
"Rule: require delivery within five days, then choose lowest cost. A costs 6 and takes 8 days; B costs 8 and takes 4; C costs 9 and takes 3."
```

### Criteria

- `option_1` (value: None): "Reject all bids"
- `option_2` (value: None): "Bid C"
- `option_3` (value: None): "Bid B"
- `option_4` (value: None): "Bid A"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0}

B is cheaper than C among bids meeting the deadline.

## contextual-choice-010

### Instruction

```json
"What's the next move under the playbook?"
```

### State

```json
"The worker alone is hung; data is safe. The playbook says restart a hung worker once before escalation, and no restart has been attempted."
```

### Criteria

- `option_1` (value: None): "Fail over to the replica"
- `option_2` (value: None): "Rotate the exposed credential"
- `option_3` (value: None): "Escalate to the duty lead"
- `option_4` (value: None): "Close the incident"
- `option_5` (value: None): "Isolate the host"
- `option_6` (value: None): "Restart the worker"
- `option_7` (value: None): "Monitor"
- `option_8` (value: None): "Roll back the release"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 1.0, 'option_7': 0.0, 'option_8': 0.0}

The first prescribed recovery step is a restart.

## contextual-choice-011

### Instruction

```json
"Which review route applies to this case under the supplied rules?"
```

### State

```json
{
  "case": "Required signature missing.",
  "rules": "Incomplete submissions first receive a document request."
}
```

### Criteria

- `option_1` (value: None): "Internal review"
- `option_2` (value: None): "Request missing documents"
- `option_3` (value: None): "Independent panel"
- `option_4` (value: None): "No review available"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0}

The completeness gate precedes substantive review.

## contextual-choice-012

### Instruction

```json
"What is the appropriate release format under this agreement?"
```

### State

```json
[
  "Agreement: authorized care team must receive identified records for follow-up.",
  "Recipient: authorized care team; transfer approved."
]
```

### Criteria

- `option_1` (value: None): "Identifiable rows"
- `option_2` (value: None): "Pseudonymized rows"
- `option_3` (value: None): "Synthetic sample"
- `option_4` (value: None): "Do not release"
- `option_5` (value: None): "Aggregated counts"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

The approved follow-up requires identifiable records.

## contextual-choice-013

### Instruction

```json
"Which item does 'that one' refer to in the final reply?"
```

### State

```json
"A points only to the green folder and asks, 'This?' B answers, 'That one, yes.' No other gesture occurs."
```

### Criteria

- `option_1` (value: None): "Yellow folder"
- `option_2` (value: None): "Red folder"
- `option_3` (value: None): "Blue folder"
- `option_4` (value: None): "Black folder"
- `option_5` (value: None): "Green folder"
- `option_6` (value: None): "White folder"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0, 'option_6': 0.0}

The sole pointing gesture identifies green.

## contextual-choice-014

### Instruction

```json
"Which use of the square best follows the council's stated priority?"
```

### State

```json
"The council prioritizes a low-noise green refuge for reading and rest. It does not want events, commerce or play equipment on this site."
```

### Criteria

- `option_1` (value: None): "Keep an open assembly space"
- `option_2` (value: None): "Install a play area"
- `option_3` (value: None): "Create a quiet garden"
- `option_4` (value: None): "Make a flood-storage basin"
- `option_5` (value: None): "Build a market hall"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0, 'option_5': 0.0}

A quiet garden matches the use and noise constraints.

## contextual-choice-015

### Instruction

```json
"Pick the route that fits the traveler's requirements."
```

### State

```json
[
  "Budget: zero. No bicycle and no free rides available.",
  "Destination: a safe ten-minute walk; no mobility limitations."
]
```

### Criteria

- `option_1` (value: None): "Cycle"
- `option_2` (value: None): "Take the bus"
- `option_3` (value: None): "Walk"
- `option_4` (value: None): "Join a carpool"
- `option_5` (value: None): "Use the accessible taxi"
- `option_6` (value: None): "Take the train"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0}

Walking is feasible and has no fare.

## contextual-choice-016

### Instruction

```json
"Which teaching response best fits this learner right now?"
```

### State

```json
"The learner cannot access the image-only worksheet with their screen reader. The task itself is at the right level."
```

### Criteria

- `option_1` (value: None): "Provide a screen-reader-friendly version"
- `option_2` (value: None): "Give a worked example"
- `option_3` (value: None): "Schedule a rest break"
- `option_4` (value: None): "Ask for the learner's reasoning"
- `option_5` (value: None): "Offer an advanced extension"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

An accessible version removes the actual barrier.

## contextual-choice-017

### Instruction

```json
"Which plan should this fictional caucus select under its decision rule?"
```

### State

```json
{
  "rule": "Choose the only plan that reduces home heat loss directly; supply expansion is secondary."
}
```

### Criteria

- `option_1` (value: None): "Tidal pilot project"
- `option_2` (value: None): "Public nuclear plant"
- `option_3` (value: None): "Rooftop solar grants"
- `option_4` (value: None): "Home insulation program"
- `option_5` (value: None): "Community wind cooperative"
- `option_6` (value: None): "Emergency diesel generators"
- `option_7` (value: None): "No new public spending"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0}

Insulation directly reduces heat loss.

## contextual-choice-018

### Instruction

```json
"Which meeting mode meets the team's constraints?"
```

### State

```json
"The team is distributed across cities but shares a working hour. Policy requires a real-time face-to-face remote review; everyone has a camera and consented."
```

### Criteria

- `option_1` (value: None): "Asynchronous written discussion"
- `option_2` (value: None): "In person"
- `option_3` (value: None): "Live video"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0}

Video meets the stipulated remote synchronous format.

## contextual-choice-019

### Instruction

```json
"Which use of the square best follows the council's stated priority?"
```

### State

```json
[
  "Priority: sheltered stalls for small traders.",
  "The other proposed uses have no trading stalls."
]
```

### Criteria

- `option_1` (value: None): "Keep an open assembly space"
- `option_2` (value: None): "Install a play area"
- `option_3` (value: None): "Create a quiet garden"
- `option_4` (value: None): "Make a flood-storage basin"
- `option_5` (value: None): "Build a market hall"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0}

The market hall provides sheltered trading space.

## contextual-choice-020

### Instruction

```json
"Which review route applies to this case under the supplied rules?"
```

### State

```json
"Rules: staff conflicts of interest bypass internal review and go directly to the independent panel. A staff conflict is established; the complete case is timely."
```

### Criteria

- `option_1` (value: None): "Internal review"
- `option_2` (value: None): "Request missing documents"
- `option_3` (value: None): "Independent panel"
- `option_4` (value: None): "No review available"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0}

The explicit bypass applies.

## contextual-choice-021

### Instruction

```json
"If the bridge were closed, which route would the dispatcher select?"
```

### State

```json
{
  "minutes": {
    "East ferry": 20,
    "North tunnel": 12,
    "South road": 30,
    "West rail": 40,
    "bridge": 4
  },
  "rule": "Choose the fastest available route; bridge removal changes nothing else."
}
```

### Criteria

- `option_1` (value: None): "Cancel the trip"
- `option_2` (value: None): "South road"
- `option_3` (value: None): "West rail"
- `option_4` (value: None): "North tunnel"
- `option_5` (value: None): "East ferry"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0}

Removing the bridge leaves the tunnel fastest.

## contextual-choice-022

### Instruction

```json
"Which bid should procurement select under the stated rule?"
```

### State

```json
[
  "Rule: certification is mandatory, with no exceptions.",
  "None of A, B or C holds the certification."
]
```

### Criteria

- `option_1` (value: None): "Reject all bids"
- `option_2` (value: None): "Bid C"
- `option_3` (value: None): "Bid B"
- `option_4` (value: None): "Bid A"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0}

Every bid fails a mandatory condition.

## contextual-choice-023

### Instruction

```json
"If the bridge were closed, which route would the dispatcher select?"
```

### State

```json
{
  "after_bridge_closure": "No remaining route is permitted.",
  "rule": "Use any available permitted route; otherwise cancel."
}
```

### Criteria

- `option_1` (value: None): "Cancel the trip"
- `option_2` (value: None): "South road"
- `option_3` (value: None): "West rail"
- `option_4` (value: None): "North tunnel"
- `option_5` (value: None): "East ferry"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

No feasible permitted route remains.

## contextual-choice-024

### Instruction

```json
"Which editorial action fits the evidence and the desk's rule?"
```

### State

```json
{
  "rule": "Publish as verified only after two independent records agree.",
  "status": "Two independent records agree; story has not been published."
}
```

### Criteria

- `option_1` (value: None): "Publish as verified"
- `option_2` (value: None): "Publish explicitly as an unverified allegation"
- `option_3` (value: None): "Publish a correction"
- `option_4` (value: None): "Hold for verification"
- `option_5` (value: None): "Retract the story"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

The verification requirement is met.

## contextual-choice-025

### Instruction

```json
"How should the facilitator resolve this decision under the group's charter?"
```

### State

```json
"The charter delegates decisions affecting only one subgroup to that subgroup. This decision affects only the night-shift subgroup."
```

### Criteria

- `option_1` (value: None): "Require unanimous consent"
- `option_2` (value: None): "Defer to the affected subgroup"
- `option_3` (value: None): "Let the rotating steward decide"
- `option_4` (value: None): "Draw lots"
- `option_5` (value: None): "Simple majority vote"
- `option_6` (value: None): "Seek external arbitration"
- `option_7` (value: None): "Run a reversible pilot"
- `option_8` (value: None): "Postpone until quorum"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0, 'option_8': 0.0}

The scope triggers delegated authority.

## contextual-choice-026

### Instruction

```json
"What's the next move under the playbook?"
```

### State

```json
"All recovery and verification checks have passed. The playbook requires closing the incident once the owner signs off; the owner just signed off."
```

### Criteria

- `option_1` (value: None): "Fail over to the replica"
- `option_2` (value: None): "Rotate the exposed credential"
- `option_3` (value: None): "Escalate to the duty lead"
- `option_4` (value: None): "Close the incident"
- `option_5` (value: None): "Isolate the host"
- `option_6` (value: None): "Restart the worker"
- `option_7` (value: None): "Monitor"
- `option_8` (value: None): "Roll back the release"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0, 'option_8': 0.0}

The final closure condition is satisfied.

## contextual-choice-027

### Instruction

```json
"If the bridge were closed, which route would the dispatcher select?"
```

### State

```json
"With the bridge removed, the only route meeting the 30-minute deadline is west rail. The dispatcher prioritizes meeting that deadline, and the load fits on rail."
```

### Criteria

- `option_1` (value: None): "Cancel the trip"
- `option_2` (value: None): "South road"
- `option_3` (value: None): "West rail"
- `option_4` (value: None): "North tunnel"
- `option_5` (value: None): "East ferry"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0, 'option_5': 0.0}

West rail uniquely meets the deadline.

## contextual-choice-028

### Instruction

```json
"Which version best matches the requested tone and meaning?"
```

### State

```json
[
  "Intent: soften the request by phrasing it as a courteous question.",
  "Required action: send the report today."
]
```

### Criteria

- `option_1` (value: None): "Would you mind sending the report today?"
- `option_2` (value: None): "The report may be sent today."
- `option_3` (value: None): "Do not send the report today."
- `option_4` (value: None): "Send the report today."
- `option_5` (value: None): "Please send the report today."

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

The courteous question preserves the requested action.

## contextual-choice-029

### Instruction

```json
"What should we serve this guest, using only their stated requirements?"
```

### State

```json
"The guest personally requires Cedar-certified food; only the chicken plate has that certification. They eat poultry and are not fasting."
```

### Criteria

- `option_1` (value: None): "Cedar-certified chicken plate"
- `option_2` (value: None): "No meal until sunset"
- `option_3` (value: None): "Fish with rice"
- `option_4` (value: None): "Lentil stew with vegetable stock"
- `option_5` (value: None): "Plain rice and sealed water"
- `option_6` (value: None): "Cheese sandwich"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0}

The certification constraint singles out the chicken plate.

## contextual-choice-030

### Instruction

```json
"Which review route applies to this case under the supplied rules?"
```

### State

```json
{
  "case": "Final panel decision already issued; no exceptions exist.",
  "rules": "No further review after a final independent-panel decision."
}
```

### Criteria

- `option_1` (value: None): "Internal review"
- `option_2` (value: None): "Request missing documents"
- `option_3` (value: None): "Independent panel"
- `option_4` (value: None): "No review available"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0}

The stated process is exhausted.

## contextual-choice-031

### Instruction

```json
"Which plan should this fictional caucus select under its decision rule?"
```

### State

```json
"The caucus demands a publicly owned nuclear generating facility. It rejects all proposed alternatives for this vote."
```

### Criteria

- `option_1` (value: None): "Tidal pilot project"
- `option_2` (value: None): "Public nuclear plant"
- `option_3` (value: None): "Rooftop solar grants"
- `option_4` (value: None): "Home insulation program"
- `option_5` (value: None): "Community wind cooperative"
- `option_6` (value: None): "Emergency diesel generators"
- `option_7` (value: None): "No new public spending"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0}

The public nuclear plant matches the explicit mandate.

## contextual-choice-032

### Instruction

```json
"Which meeting mode meets the team's constraints?"
```

### State

```json
{
  "constraints": "No overlapping working hours; everyone has text access."
}
```

### Criteria

- `option_1` (value: None): "Asynchronous written discussion"
- `option_2` (value: None): "In person"
- `option_3` (value: None): "Live video"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0}

Asynchronous writing needs no shared meeting time.

## contextual-choice-033

### Instruction

```json
"What should we serve this guest, using only their stated requirements?"
```

### State

```json
{
  "requirements": "No animal products; wants a substantial legume dish; not fasting."
}
```

### Criteria

- `option_1` (value: None): "Cedar-certified chicken plate"
- `option_2` (value: None): "No meal until sunset"
- `option_3` (value: None): "Fish with rice"
- `option_4` (value: None): "Lentil stew with vegetable stock"
- `option_5` (value: None): "Plain rice and sealed water"
- `option_6` (value: None): "Cheese sandwich"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0, 'option_6': 0.0}

The lentil dish is the only substantial legume option without animal products.

## contextual-choice-034

### Instruction

```json
"Which subscription best matches the stated requirement?"
```

### State

```json
[
  "The required resource is freely available without an account.",
  "The user wants to avoid any unnecessary payment; paid plans add no needed feature."
]
```

### Criteria

- `option_1` (value: None): "Monthly plan"
- `option_2` (value: None): "No subscription"
- `option_3` (value: None): "Annual plan"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0}

No subscription meets the need without payment.

## contextual-choice-035

### Instruction

```json
"Pick the route that fits the traveler's requirements."
```

### State

```json
{
  "need": "Work at a table with power during the trip.",
  "services": "Only the train provides a table and power."
}
```

### Criteria

- `option_1` (value: None): "Cycle"
- `option_2` (value: None): "Take the bus"
- `option_3` (value: None): "Walk"
- `option_4` (value: None): "Join a carpool"
- `option_5` (value: None): "Use the accessible taxi"
- `option_6` (value: None): "Take the train"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 1.0}

Only the train provides both required facilities.

## contextual-choice-036

### Instruction

```json
"If the bridge were closed, which route would the dispatcher select?"
```

### State

```json
[
  "After bridge closure, the tunnel and road are unavailable; rail does not carry this load.",
  "The ferry is operating and can carry it; any feasible trip is preferred to cancellation."
]
```

### Criteria

- `option_1` (value: None): "Cancel the trip"
- `option_2` (value: None): "South road"
- `option_3` (value: None): "West rail"
- `option_4` (value: None): "North tunnel"
- `option_5` (value: None): "East ferry"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0}

The ferry is the only feasible remaining mode.

## contextual-choice-037

### Instruction

```json
"Which editorial action fits the evidence and the desk's rule?"
```

### State

```json
{
  "rule": "Do not publish anonymous single-source allegations without corroboration.",
  "status": "Unpublished, one anonymous source, no corroboration."
}
```

### Criteria

- `option_1` (value: None): "Publish as verified"
- `option_2` (value: None): "Publish explicitly as an unverified allegation"
- `option_3` (value: None): "Publish a correction"
- `option_4` (value: None): "Hold for verification"
- `option_5` (value: None): "Retract the story"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0}

Holding is required until verification.

## contextual-choice-038

### Instruction

```json
"Pick the route that fits the traveler's requirements."
```

### State

```json
{
  "availability": "Only the accessible taxi is step-free door-to-door.",
  "needs": "Step-free door-to-door travel; no transfers."
}
```

### Criteria

- `option_1` (value: None): "Cycle"
- `option_2` (value: None): "Take the bus"
- `option_3` (value: None): "Walk"
- `option_4` (value: None): "Join a carpool"
- `option_5` (value: None): "Use the accessible taxi"
- `option_6` (value: None): "Take the train"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0, 'option_6': 0.0}

The taxi uniquely meets the accessibility and transfer constraints.

## contextual-choice-039

### Instruction

```json
"What is the appropriate release format under this agreement?"
```

### State

```json
{
  "agreement": "Share row-level research records with stable random IDs and no direct identifiers.",
  "recipient": "approved researcher"
}
```

### Criteria

- `option_1` (value: None): "Identifiable rows"
- `option_2` (value: None): "Pseudonymized rows"
- `option_3` (value: None): "Synthetic sample"
- `option_4` (value: None): "Do not release"
- `option_5` (value: None): "Aggregated counts"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

Pseudonymized rows meet the stipulated format.

## contextual-choice-040

### Instruction

```json
"Which version best matches the requested tone and meaning?"
```

### State

```json
"The intended message forbids sending the report today. It must not be softened into a suggestion or permission."
```

### Criteria

- `option_1` (value: None): "Would you mind sending the report today?"
- `option_2` (value: None): "The report may be sent today."
- `option_3` (value: None): "Do not send the report today."
- `option_4` (value: None): "Send the report today."
- `option_5` (value: None): "Please send the report today."

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0, 'option_5': 0.0}

Only the negative imperative preserves the prohibition.

## contextual-choice-041

### Instruction

```json
"Which editorial action fits the evidence and the desk's rule?"
```

### State

```json
"The desk permits public-interest allegations before verification only with explicit uncertainty labeling. The allegation meets that policy; the editor has decided it should run now."
```

### Criteria

- `option_1` (value: None): "Publish as verified"
- `option_2` (value: None): "Publish explicitly as an unverified allegation"
- `option_3` (value: None): "Publish a correction"
- `option_4` (value: None): "Hold for verification"
- `option_5` (value: None): "Retract the story"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

The uncertainty-labeled publication is the allowed immediate action.

## contextual-choice-042

### Instruction

```json
"Which event format respects the organizer's stated commitments?"
```

### State

```json
"The organizer wants to learn the building's history without participating in or observing a rite. They want an in-person educational talk, not a discussion."
```

### Criteria

- `option_1` (value: None): "Interfaith dialogue circle"
- `option_2` (value: None): "Silent observation of a rite"
- `option_3` (value: None): "Postpone the event"
- `option_4` (value: None): "Outdoor communal singing"
- `option_5` (value: None): "Participatory prayer"
- `option_6` (value: None): "Secular history lecture"
- `option_7` (value: None): "Private reflection at home"
- `option_8` (value: None): "Online text discussion"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 1.0, 'option_7': 0.0, 'option_8': 0.0}

The secular lecture matches the educational format and exclusions.

## contextual-choice-043

### Instruction

```json
"What should we serve this guest, using only their stated requirements?"
```

### State

```json
"The guest avoids meat and fish but eats dairy. They request a cold handheld lunch and dislike legumes. Only the sandwich is cold and handheld."
```

### Criteria

- `option_1` (value: None): "Cedar-certified chicken plate"
- `option_2` (value: None): "No meal until sunset"
- `option_3` (value: None): "Fish with rice"
- `option_4` (value: None): "Lentil stew with vegetable stock"
- `option_5` (value: None): "Plain rice and sealed water"
- `option_6` (value: None): "Cheese sandwich"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 1.0}

The sandwich meets the dietary and serving preferences.

## contextual-choice-044

### Instruction

```json
"Which teaching response best fits this learner right now?"
```

### State

```json
"The learner is rested and engaged but has never seen the procedure. They explicitly ask to see one complete demonstration before trying."
```

### Criteria

- `option_1` (value: None): "Provide a screen-reader-friendly version"
- `option_2` (value: None): "Give a worked example"
- `option_3` (value: None): "Schedule a rest break"
- `option_4` (value: None): "Ask for the learner's reasoning"
- `option_5` (value: None): "Offer an advanced extension"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

A worked example provides the requested demonstration.

## contextual-choice-045

### Instruction

```json
"Which use of the square best follows the council's stated priority?"
```

### State

```json
{
  "priority": "Protect space for public demonstrations; no permanent structures.",
  "site": "Only large central gathering area."
}
```

### Criteria

- `option_1` (value: None): "Keep an open assembly space"
- `option_2` (value: None): "Install a play area"
- `option_3` (value: None): "Create a quiet garden"
- `option_4` (value: None): "Make a flood-storage basin"
- `option_5` (value: None): "Build a market hall"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

Open assembly space serves the explicit civic priority.

## contextual-choice-046

### Instruction

```json
"Which review route applies to this case under the supplied rules?"
```

### State

```json
[
  "Rules: after internal review, the independent panel is the final route.",
  "Internal review finished; the timely appeal file is complete."
]
```

### Criteria

- `option_1` (value: None): "Internal review"
- `option_2` (value: None): "Request missing documents"
- `option_3` (value: None): "Independent panel"
- `option_4` (value: None): "No review available"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0}

The completed first stage unlocks the independent panel.

## contextual-choice-047

### Instruction

```json
"Which subscription best matches the stated requirement?"
```

### State

```json
"The user requires a feature available only on the annual plan and has approved its cost. Neither the monthly plan nor free access includes it."
```

### Criteria

- `option_1` (value: None): "Monthly plan"
- `option_2` (value: None): "No subscription"
- `option_3` (value: None): "Annual plan"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0}

Only the annual plan covers the required feature.

## contextual-choice-048

### Instruction

```json
"Which meeting mode meets the team's constraints?"
```

### State

```json
{
  "constraints": "Audio and video are unavailable this week; accessible text is available.",
  "urgency": "Replies within two days are acceptable."
}
```

### Criteria

- `option_1` (value: None): "Asynchronous written discussion"
- `option_2` (value: None): "In person"
- `option_3` (value: None): "Live video"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0}

Written asynchronous discussion remains feasible.

## contextual-choice-049

### Instruction

```json
"Which review route applies to this case under the supplied rules?"
```

### State

```json
"Rules: initial decisions go to internal review. This is the first challenge to an initial decision and the file is complete."
```

### Criteria

- `option_1` (value: None): "Internal review"
- `option_2` (value: None): "Request missing documents"
- `option_3` (value: None): "Independent panel"
- `option_4` (value: None): "No review available"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0}

The initial challenge uses internal review.

## contextual-choice-050

### Instruction

```json
"Which plan should this fictional caucus select under its decision rule?"
```

### State

```json
"The caucus prioritizes resident ownership of generation. Only the wind cooperative gives residents ownership; all other generation projects are privately or nationally owned."
```

### Criteria

- `option_1` (value: None): "Tidal pilot project"
- `option_2` (value: None): "Public nuclear plant"
- `option_3` (value: None): "Rooftop solar grants"
- `option_4` (value: None): "Home insulation program"
- `option_5` (value: None): "Community wind cooperative"
- `option_6` (value: None): "Emergency diesel generators"
- `option_7` (value: None): "No new public spending"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0, 'option_6': 0.0, 'option_7': 0.0}

The cooperative uniquely satisfies ownership priority.

## contextual-choice-051

### Instruction

```json
"Which teaching response best fits this learner right now?"
```

### State

```json
{
  "learner": "Understands the concept and finishes every basic task accurately.",
  "request": "A harder challenge, please."
}
```

### Criteria

- `option_1` (value: None): "Provide a screen-reader-friendly version"
- `option_2` (value: None): "Give a worked example"
- `option_3` (value: None): "Schedule a rest break"
- `option_4` (value: None): "Ask for the learner's reasoning"
- `option_5` (value: None): "Offer an advanced extension"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0}

An extension addresses demonstrated mastery and the request.

## contextual-choice-052

### Instruction

```json
"Where should this council allocate the next unit of water?"
```

### State

```json
"The council's adopted rule gives drinking water first priority until every household has the minimum. Some households remain below that minimum; hospital reserves are adequate."
```

### Criteria

- `option_1` (value: None): "Firefighting reserve"
- `option_2` (value: None): "Hospital reserve"
- `option_3` (value: None): "River habitat"
- `option_4` (value: None): "Industrial cooling"
- `option_5` (value: None): "Food crops"
- `option_6` (value: None): "Household drinking supply"
- `option_7` (value: None): "Public fountains"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 1.0, 'option_7': 0.0}

Unmet household minimum has first priority.

## contextual-choice-053

### Instruction

```json
"What should we serve this guest, using only their stated requirements?"
```

### State

```json
{
  "requirements": "For today's observance, only plain rice and sealed water; no additions."
}
```

### Criteria

- `option_1` (value: None): "Cedar-certified chicken plate"
- `option_2` (value: None): "No meal until sunset"
- `option_3` (value: None): "Fish with rice"
- `option_4` (value: None): "Lentil stew with vegetable stock"
- `option_5` (value: None): "Plain rice and sealed water"
- `option_6` (value: None): "Cheese sandwich"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0, 'option_6': 0.0}

Only that option matches the exact temporary restriction.

## contextual-choice-054

### Instruction

```json
"How should the facilitator resolve this decision under the group's charter?"
```

### State

```json
{
  "attendance": "Below quorum.",
  "charter": "Without quorum, no decision method may proceed."
}
```

### Criteria

- `option_1` (value: None): "Require unanimous consent"
- `option_2` (value: None): "Defer to the affected subgroup"
- `option_3` (value: None): "Let the rotating steward decide"
- `option_4` (value: None): "Draw lots"
- `option_5` (value: None): "Simple majority vote"
- `option_6` (value: None): "Seek external arbitration"
- `option_7` (value: None): "Run a reversible pilot"
- `option_8` (value: None): "Postpone until quorum"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0, 'option_8': 1.0}

The quorum prerequisite blocks other methods.

## contextual-choice-055

### Instruction

```json
"What's the next move under the playbook?"
```

### State

```json
[
  "Playbook: isolate any host with confirmed active malicious execution.",
  "This host has confirmed active malicious execution."
]
```

### Criteria

- `option_1` (value: None): "Fail over to the replica"
- `option_2` (value: None): "Rotate the exposed credential"
- `option_3` (value: None): "Escalate to the duty lead"
- `option_4` (value: None): "Close the incident"
- `option_5` (value: None): "Isolate the host"
- `option_6` (value: None): "Restart the worker"
- `option_7` (value: None): "Monitor"
- `option_8` (value: None): "Roll back the release"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0, 'option_6': 0.0, 'option_7': 0.0, 'option_8': 0.0}

The containment trigger calls for isolation.

## contextual-choice-056

### Instruction

```json
"Which bid should procurement select under the stated rule?"
```

### State

```json
{
  "rule": "Choose highest verified durability; cost is irrelevant.",
  "verified_years": {
    "A": 2,
    "B": 3,
    "C": 7
  }
}
```

### Criteria

- `option_1` (value: None): "Reject all bids"
- `option_2` (value: None): "Bid C"
- `option_3` (value: None): "Bid B"
- `option_4` (value: None): "Bid A"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0}

C has the greatest verified durability.

## contextual-choice-057

### Instruction

```json
"Pick the route that fits the traveler's requirements."
```

### State

```json
"The traveler wants exercise, owns a bicycle and must cover 12 km in 45 minutes. Walking takes 150 minutes; cycling takes 40. They exclude motorized travel."
```

### Criteria

- `option_1` (value: None): "Cycle"
- `option_2` (value: None): "Take the bus"
- `option_3` (value: None): "Walk"
- `option_4` (value: None): "Join a carpool"
- `option_5` (value: None): "Use the accessible taxi"
- `option_6` (value: None): "Take the train"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0}

Cycling is the only permitted mode within the time limit.

## contextual-choice-058

### Instruction

```json
"Which subscription best matches the stated requirement?"
```

### State

```json
{
  "need": "Access for one month only.",
  "prices": {
    "annual": 80,
    "monthly": 10
  },
  "rule": "Buy the cheapest plan covering the need."
}
```

### Criteria

- `option_1` (value: None): "Monthly plan"
- `option_2` (value: None): "No subscription"
- `option_3` (value: None): "Annual plan"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0}

One monthly payment costs less than annual coverage.

## contextual-choice-059

### Instruction

```json
"Which item does 'that one' refer to in the final reply?"
```

### State

```json
[
  {
    "speaker": "A",
    "text": "Do you mean the blue folder?"
  },
  {
    "speaker": "B",
    "text": "Yes, that one."
  }
]
```

### Criteria

- `option_1` (value: None): "Yellow folder"
- `option_2` (value: None): "Red folder"
- `option_3` (value: None): "Blue folder"
- `option_4` (value: None): "Black folder"
- `option_5` (value: None): "Green folder"
- `option_6` (value: None): "White folder"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0}

The reply confirms the immediately named blue folder.

## contextual-choice-060

### Instruction

```json
"Which treatment follows the conservator's stated constraints?"
```

### State

```json
"The object's coating is unidentified. Protocol forbids any treatment until a specialist identifies it, including tests and brushing."
```

### Criteria

- `option_1` (value: None): "Solvent wipe"
- `option_2` (value: None): "Light exposure test"
- `option_3` (value: None): "Water rinse"
- `option_4` (value: None): "Controlled humidification"
- `option_5` (value: None): "Leave untouched and consult specialist"
- `option_6` (value: None): "Freeze treatment"
- `option_7` (value: None): "Dry brush"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0, 'option_6': 0.0, 'option_7': 0.0}

Consultation without treatment is required.

## contextual-choice-061

### Instruction

```json
"Should the application enable this sharing feature now?"
```

### State

```json
"Permission was for location sharing with a named friend only. This feature sends location to advertisers, requiring separate consent that was not given."
```

### Criteria

- `option_1` (value: None): "Enable sharing"
- `option_2` (value: None): "Keep sharing disabled"

Target: {'option_1': 0.0, 'option_2': 1.0}

The existing permission does not cover this recipient.

## contextual-choice-062

### Instruction

```json
"Where should this council allocate the next unit of water?"
```

### State

```json
"All protected reserves and ecological minimums are met. The council's next priority is staple food production; crop irrigation is short, while industrial cooling is optional."
```

### Criteria

- `option_1` (value: None): "Firefighting reserve"
- `option_2` (value: None): "Hospital reserve"
- `option_3` (value: None): "River habitat"
- `option_4` (value: None): "Industrial cooling"
- `option_5` (value: None): "Food crops"
- `option_6` (value: None): "Household drinking supply"
- `option_7` (value: None): "Public fountains"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0, 'option_6': 0.0, 'option_7': 0.0}

Food crops are the highest remaining priority.

## contextual-choice-063

### Instruction

```json
"How should the facilitator resolve this decision under the group's charter?"
```

### State

```json
"The charter requires external arbitration for unresolved conflicts between two teams. Both teams completed internal mediation without agreement."
```

### Criteria

- `option_1` (value: None): "Require unanimous consent"
- `option_2` (value: None): "Defer to the affected subgroup"
- `option_3` (value: None): "Let the rotating steward decide"
- `option_4` (value: None): "Draw lots"
- `option_5` (value: None): "Simple majority vote"
- `option_6` (value: None): "Seek external arbitration"
- `option_7` (value: None): "Run a reversible pilot"
- `option_8` (value: None): "Postpone until quorum"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 1.0, 'option_7': 0.0, 'option_8': 0.0}

The failed internal process triggers arbitration.

## contextual-choice-064

### Instruction

```json
"Where should this council allocate the next unit of water?"
```

### State

```json
{
  "priority_rule": "Meet emergency fire reserve before all other uses when wildfire alert is active.",
  "status": "Wildfire alert active; fire reserve below minimum."
}
```

### Criteria

- `option_1` (value: None): "Firefighting reserve"
- `option_2` (value: None): "Hospital reserve"
- `option_3` (value: None): "River habitat"
- `option_4` (value: None): "Industrial cooling"
- `option_5` (value: None): "Food crops"
- `option_6` (value: None): "Household drinking supply"
- `option_7` (value: None): "Public fountains"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0}

The emergency priority applies.

## contextual-choice-065

### Instruction

```json
"Which treatment follows the conservator's stated constraints?"
```

### State

```json
"Protocol calls for a light exposure test on this approved sacrificial sample before deciding storage. Cleaning or other treatment would invalidate that test."
```

### Criteria

- `option_1` (value: None): "Solvent wipe"
- `option_2` (value: None): "Light exposure test"
- `option_3` (value: None): "Water rinse"
- `option_4` (value: None): "Controlled humidification"
- `option_5` (value: None): "Leave untouched and consult specialist"
- `option_6` (value: None): "Freeze treatment"
- `option_7` (value: None): "Dry brush"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0}

The test is the required next action.

## contextual-choice-066

### Instruction

```json
"Should the application enable this sharing feature now?"
```

### State

```json
[
  "The user opted in last week.",
  "Today they revoked permission. Revocations take effect immediately."
]
```

### Criteria

- `option_1` (value: None): "Enable sharing"
- `option_2` (value: None): "Keep sharing disabled"

Target: {'option_1': 0.0, 'option_2': 1.0}

Current revocation overrides past permission.

## contextual-choice-067

### Instruction

```json
"Which use of the square best follows the council's stated priority?"
```

### State

```json
{
  "priority": "A place for children's active outdoor play.",
  "residents": "Many families; nearby parks prohibit play equipment."
}
```

### Criteria

- `option_1` (value: None): "Keep an open assembly space"
- `option_2` (value: None): "Install a play area"
- `option_3` (value: None): "Create a quiet garden"
- `option_4` (value: None): "Make a flood-storage basin"
- `option_5` (value: None): "Build a market hall"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

The play area fills the specified need.

## contextual-choice-068

### Instruction

```json
"Which subscription best matches the stated requirement?"
```

### State

```json
{
  "need": "Two months of access, then stop.",
  "prices": {
    "annual": 80,
    "monthly": 10
  },
  "rule": "Minimum total cost."
}
```

### Criteria

- `option_1` (value: None): "Monthly plan"
- `option_2` (value: None): "No subscription"
- `option_3` (value: None): "Annual plan"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0}

Two monthly payments total 20, below 80.

## contextual-choice-069

### Instruction

```json
"Which plan should this fictional caucus select under its decision rule?"
```

### State

```json
[
  "Binding mandate: authorize no additional public expenditure this year.",
  "Every construction or grant option requires additional public expenditure."
]
```

### Criteria

- `option_1` (value: None): "Tidal pilot project"
- `option_2` (value: None): "Public nuclear plant"
- `option_3` (value: None): "Rooftop solar grants"
- `option_4` (value: None): "Home insulation program"
- `option_5` (value: None): "Community wind cooperative"
- `option_6` (value: None): "Emergency diesel generators"
- `option_7` (value: None): "No new public spending"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 1.0}

Only no new spending satisfies the mandate.

## contextual-choice-070

### Instruction

```json
"Where should this council allocate the next unit of water?"
```

### State

```json
[
  "Rule: after human emergency needs are met, prevent a river habitat collapse before supporting production.",
  "Emergency needs met; river at collapse threshold."
]
```

### Criteria

- `option_1` (value: None): "Firefighting reserve"
- `option_2` (value: None): "Hospital reserve"
- `option_3` (value: None): "River habitat"
- `option_4` (value: None): "Industrial cooling"
- `option_5` (value: None): "Food crops"
- `option_6` (value: None): "Household drinking supply"
- `option_7` (value: None): "Public fountains"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0}

The stated ecological priority is now active.

## contextual-choice-071

### Instruction

```json
"Which item does 'that one' refer to in the final reply?"
```

### State

```json
[
  "A: Not the yellow folder; I need the white folder.",
  "B: I'll fetch that one."
]
```

### Criteria

- `option_1` (value: None): "Yellow folder"
- `option_2` (value: None): "Red folder"
- `option_3` (value: None): "Blue folder"
- `option_4` (value: None): "Black folder"
- `option_5` (value: None): "Green folder"
- `option_6` (value: None): "White folder"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 1.0}

The affirmative requested referent is white.

## contextual-choice-072

### Instruction

```json
"Which editorial action fits the evidence and the desk's rule?"
```

### State

```json
[
  "The published story's central claim is true, but its date is wrong.",
  "Desk policy: correct isolated factual details while keeping the story."
]
```

### Criteria

- `option_1` (value: None): "Publish as verified"
- `option_2` (value: None): "Publish explicitly as an unverified allegation"
- `option_3` (value: None): "Publish a correction"
- `option_4` (value: None): "Hold for verification"
- `option_5` (value: None): "Retract the story"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0, 'option_5': 0.0}

A correction addresses the isolated date error.

## contextual-choice-073

### Instruction

```json
"Which editorial action fits the evidence and the desk's rule?"
```

### State

```json
"The story is already published. Its central claim is established false. Desk rule: retract when the central claim is false, rather than merely correcting details."
```

### Criteria

- `option_1` (value: None): "Publish as verified"
- `option_2` (value: None): "Publish explicitly as an unverified allegation"
- `option_3` (value: None): "Publish a correction"
- `option_4` (value: None): "Hold for verification"
- `option_5` (value: None): "Retract the story"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0}

The central falsehood triggers retraction.

## contextual-choice-074

### Instruction

```json
"Which meeting mode meets the team's constraints?"
```

### State

```json
[
  "The session requires handling a physical prototype that cannot leave the lab.",
  "All participants can attend the lab."
]
```

### Criteria

- `option_1` (value: None): "Asynchronous written discussion"
- `option_2` (value: None): "In person"
- `option_3` (value: None): "Live video"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0}

In-person attendance enables physical handling.

## contextual-choice-075

### Instruction

```json
"Which treatment follows the conservator's stated constraints?"
```

### State

```json
[
  "Protocol: treat this verified pest infestation by freezing; other listed methods are ineffective.",
  "The object is certified safe to freeze."
]
```

### Criteria

- `option_1` (value: None): "Solvent wipe"
- `option_2` (value: None): "Light exposure test"
- `option_3` (value: None): "Water rinse"
- `option_4` (value: None): "Controlled humidification"
- `option_5` (value: None): "Leave untouched and consult specialist"
- `option_6` (value: None): "Freeze treatment"
- `option_7` (value: None): "Dry brush"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 1.0, 'option_7': 0.0}

Freezing is both prescribed and safe under the supplied facts.

## contextual-choice-076

### Instruction

```json
"Which item does 'that one' refer to in the final reply?"
```

### State

```json
"A: The red folder is the only signed copy. B: Then send that one. The other colors are unsigned."
```

### Criteria

- `option_1` (value: None): "Yellow folder"
- `option_2` (value: None): "Red folder"
- `option_3` (value: None): "Blue folder"
- `option_4` (value: None): "Black folder"
- `option_5` (value: None): "Green folder"
- `option_6` (value: None): "White folder"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0}

The signed-copy description uniquely identifies red.

## contextual-choice-077

### Instruction

```json
"What's the next move under the playbook?"
```

### State

```json
{
  "event": "A valid credential was published.",
  "playbook": "Confirmed credential exposure requires immediate rotation before any other action."
}
```

### Criteria

- `option_1` (value: None): "Fail over to the replica"
- `option_2` (value: None): "Rotate the exposed credential"
- `option_3` (value: None): "Escalate to the duty lead"
- `option_4` (value: None): "Close the incident"
- `option_5` (value: None): "Isolate the host"
- `option_6` (value: None): "Restart the worker"
- `option_7` (value: None): "Monitor"
- `option_8` (value: None): "Roll back the release"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0, 'option_8': 0.0}

The exposure rule explicitly sets the next action.

## contextual-choice-078

### Instruction

```json
"What is the appropriate release format under this agreement?"
```

### State

```json
"The recipient needs a public schema demonstration. The agreement permits only fabricated records containing no real person's data, including no aggregates of real records."
```

### Criteria

- `option_1` (value: None): "Identifiable rows"
- `option_2` (value: None): "Pseudonymized rows"
- `option_3` (value: None): "Synthetic sample"
- `option_4` (value: None): "Do not release"
- `option_5` (value: None): "Aggregated counts"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0, 'option_5': 0.0}

A synthetic sample is the only allowed demonstration.

## contextual-choice-079

### Instruction

```json
"Which use of the square best follows the council's stated priority?"
```

### State

```json
"The council's sole priority for this site is temporary stormwater retention. All options are feasible; only the basin stores floodwater."
```

### Criteria

- `option_1` (value: None): "Keep an open assembly space"
- `option_2` (value: None): "Install a play area"
- `option_3` (value: None): "Create a quiet garden"
- `option_4` (value: None): "Make a flood-storage basin"
- `option_5` (value: None): "Build a market hall"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0}

The basin directly meets the chosen objective.

## contextual-choice-080

### Instruction

```json
"Which bid should procurement select under the stated rule?"
```

### State

```json
{
  "bids": {
    "A": {
      "compliant": true,
      "cost": 7
    },
    "B": {
      "compliant": false,
      "cost": 5
    },
    "C": {
      "compliant": true,
      "cost": 9
    }
  },
  "rule": "Lowest cost among compliant bids."
}
```

### Criteria

- `option_1` (value: None): "Reject all bids"
- `option_2` (value: None): "Bid C"
- `option_3` (value: None): "Bid B"
- `option_4` (value: None): "Bid A"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0}

A is cheapest among compliant bids.

## contextual-choice-081

### Instruction

```json
"What is the appropriate release format under this agreement?"
```

### State

```json
"The agreement forbids any release, including synthetic derivatives, until consent is renewed. Consent has not been renewed."
```

### Criteria

- `option_1` (value: None): "Identifiable rows"
- `option_2` (value: None): "Pseudonymized rows"
- `option_3` (value: None): "Synthetic sample"
- `option_4` (value: None): "Do not release"
- `option_5` (value: None): "Aggregated counts"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0}

Every release format is currently prohibited.

## contextual-choice-082

### Instruction

```json
"If the bridge were closed, which route would the dispatcher select?"
```

### State

```json
"The load can travel only by road. Without the bridge, the south road is the only road route that can carry it; the tunnel is too low. Nothing else changes."
```

### Criteria

- `option_1` (value: None): "Cancel the trip"
- `option_2` (value: None): "South road"
- `option_3` (value: None): "West rail"
- `option_4` (value: None): "North tunnel"
- `option_5` (value: None): "East ferry"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0}

Only the south road remains compatible.

## contextual-choice-083

### Instruction

```json
"Which event format respects the organizer's stated commitments?"
```

### State

```json
[
  "Goal: reciprocal conversation among members of different faiths.",
  "Requirement: participants speak and listen in a circle at the venue."
]
```

### Criteria

- `option_1` (value: None): "Interfaith dialogue circle"
- `option_2` (value: None): "Silent observation of a rite"
- `option_3` (value: None): "Postpone the event"
- `option_4` (value: None): "Outdoor communal singing"
- `option_5` (value: None): "Participatory prayer"
- `option_6` (value: None): "Secular history lecture"
- `option_7` (value: None): "Private reflection at home"
- `option_8` (value: None): "Online text discussion"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0, 'option_8': 0.0}

The dialogue circle supplies the requested reciprocal format.

## contextual-choice-084

### Instruction

```json
"Which item does 'that one' refer to in the final reply?"
```

### State

```json
{
  "dialogue": [
    "A: Red or black?",
    "B: The latter.",
    "A: That one is locked."
  ],
  "rule": "The final reply keeps B's selected referent."
}
```

### Criteria

- `option_1` (value: None): "Yellow folder"
- `option_2` (value: None): "Red folder"
- `option_3` (value: None): "Blue folder"
- `option_4` (value: None): "Black folder"
- `option_5` (value: None): "Green folder"
- `option_6` (value: None): "White folder"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0, 'option_6': 0.0}

The latter of red and black is black.

## contextual-choice-085

### Instruction

```json
"What's the next move under the playbook?"
```

### State

```json
{
  "event": "Both conditions are confirmed.",
  "playbook": "Roll back if errors start immediately after deployment and the previous release is healthy."
}
```

### Criteria

- `option_1` (value: None): "Fail over to the replica"
- `option_2` (value: None): "Rotate the exposed credential"
- `option_3` (value: None): "Escalate to the duty lead"
- `option_4` (value: None): "Close the incident"
- `option_5` (value: None): "Isolate the host"
- `option_6` (value: None): "Restart the worker"
- `option_7` (value: None): "Monitor"
- `option_8` (value: None): "Roll back the release"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0, 'option_8': 1.0}

The rollback conditions are met.

## contextual-choice-086

### Instruction

```json
"How should the facilitator resolve this decision under the group's charter?"
```

### State

```json
[
  "Charter: reversible uncertain changes must first be piloted.",
  "The change is reversible and its effects are uncertain."
]
```

### Criteria

- `option_1` (value: None): "Require unanimous consent"
- `option_2` (value: None): "Defer to the affected subgroup"
- `option_3` (value: None): "Let the rotating steward decide"
- `option_4` (value: None): "Draw lots"
- `option_5` (value: None): "Simple majority vote"
- `option_6` (value: None): "Seek external arbitration"
- `option_7` (value: None): "Run a reversible pilot"
- `option_8` (value: None): "Postpone until quorum"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 1.0, 'option_8': 0.0}

The pilot rule applies.

## contextual-choice-087

### Instruction

```json
"Which teaching response best fits this learner right now?"
```

### State

```json
[
  "The learner gives an unexpected answer without showing any work.",
  "Teacher's immediate goal: diagnose the reasoning before intervening."
]
```

### Criteria

- `option_1` (value: None): "Provide a screen-reader-friendly version"
- `option_2` (value: None): "Give a worked example"
- `option_3` (value: None): "Schedule a rest break"
- `option_4` (value: None): "Ask for the learner's reasoning"
- `option_5` (value: None): "Offer an advanced extension"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0}

Asking for reasoning serves the diagnostic goal.

## contextual-choice-088

### Instruction

```json
"Which subscription best matches the stated requirement?"
```

### State

```json
"Access is needed for all twelve months. Monthly costs 10, annual costs 80; no cancellation is expected. Choose minimum total cost."
```

### Criteria

- `option_1` (value: None): "Monthly plan"
- `option_2` (value: None): "No subscription"
- `option_3` (value: None): "Annual plan"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0}

Annual coverage costs 80 rather than 120.

## contextual-choice-089

### Instruction

```json
"Which treatment follows the conservator's stated constraints?"
```

### State

```json
{
  "constraints": "Flatten the curled sheet by controlled humidification; immersion, solvents and brushing are prohibited."
}
```

### Criteria

- `option_1` (value: None): "Solvent wipe"
- `option_2` (value: None): "Light exposure test"
- `option_3` (value: None): "Water rinse"
- `option_4` (value: None): "Controlled humidification"
- `option_5` (value: None): "Leave untouched and consult specialist"
- `option_6` (value: None): "Freeze treatment"
- `option_7` (value: None): "Dry brush"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0}

The named flattening procedure matches the constraints.

## contextual-choice-090

### Instruction

```json
"Which event format respects the organizer's stated commitments?"
```

### State

```json
{
  "commitment": "Observe the rite respectfully without joining it; attendance at the site matters."
}
```

### Criteria

- `option_1` (value: None): "Interfaith dialogue circle"
- `option_2` (value: None): "Silent observation of a rite"
- `option_3` (value: None): "Postpone the event"
- `option_4` (value: None): "Outdoor communal singing"
- `option_5` (value: None): "Participatory prayer"
- `option_6` (value: None): "Secular history lecture"
- `option_7` (value: None): "Private reflection at home"
- `option_8` (value: None): "Online text discussion"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0, 'option_8': 0.0}

Silent observation meets both requirements.

## contextual-choice-091

### Instruction

```json
"Which meeting mode meets the team's constraints?"
```

### State

```json
"Everyone can meet at noon, cannot travel, and needs live visual demonstration. Video is available to everyone."
```

### Criteria

- `option_1` (value: None): "Asynchronous written discussion"
- `option_2` (value: None): "In person"
- `option_3` (value: None): "Live video"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0}

Live video satisfies real-time visual access without travel.

## contextual-choice-092

### Instruction

```json
"Which version best matches the requested tone and meaning?"
```

### State

```json
{
  "intent": "A polite but direct request to send today; avoid a question form."
}
```

### Criteria

- `option_1` (value: None): "Would you mind sending the report today?"
- `option_2` (value: None): "The report may be sent today."
- `option_3` (value: None): "Do not send the report today."
- `option_4` (value: None): "Send the report today."
- `option_5` (value: None): "Please send the report today."

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 1.0}

Please makes the direct request polite without a question.

## contextual-choice-093

### Instruction

```json
"Which version best matches the requested tone and meaning?"
```

### State

```json
"The speaker is issuing a terse imperative, without politeness markers, requiring delivery today."
```

### Criteria

- `option_1` (value: None): "Would you mind sending the report today?"
- `option_2` (value: None): "The report may be sent today."
- `option_3` (value: None): "Do not send the report today."
- `option_4` (value: None): "Send the report today."
- `option_5` (value: None): "Please send the report today."

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 1.0, 'option_5': 0.0}

The bare imperative matches the specified tone.

## contextual-choice-094

### Instruction

```json
"Pick the route that fits the traveler's requirements."
```

### State

```json
"The traveler has a bus-only pass, no cash and needs to arrive in 20 minutes. Walking takes 60; the bus takes 15. They have no bicycle and no carpool invitation."
```

### Criteria

- `option_1` (value: None): "Cycle"
- `option_2` (value: None): "Take the bus"
- `option_3` (value: None): "Walk"
- `option_4` (value: None): "Join a carpool"
- `option_5` (value: None): "Use the accessible taxi"
- `option_6` (value: None): "Take the train"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0}

The bus is both affordable and fast enough.

## contextual-choice-095

### Instruction

```json
"Which treatment follows the conservator's stated constraints?"
```

### State

```json
{
  "constraints": "Loose surface dust only; no liquid, freezing or light tests. Dry brushing explicitly approved."
}
```

### Criteria

- `option_1` (value: None): "Solvent wipe"
- `option_2` (value: None): "Light exposure test"
- `option_3` (value: None): "Water rinse"
- `option_4` (value: None): "Controlled humidification"
- `option_5` (value: None): "Leave untouched and consult specialist"
- `option_6` (value: None): "Freeze treatment"
- `option_7` (value: None): "Dry brush"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 1.0}

Dry brushing is the expressly approved compatible action.

## contextual-choice-096

### Instruction

```json
"Which bid should procurement select under the stated rule?"
```

### State

```json
"Rule: award to the only employee-owned bidder if one exists. A and C are investor-owned; B is employee-owned. All bids meet the minimum specifications."
```

### Criteria

- `option_1` (value: None): "Reject all bids"
- `option_2` (value: None): "Bid C"
- `option_3` (value: None): "Bid B"
- `option_4` (value: None): "Bid A"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0}

The ownership rule singles out B.

## contextual-choice-097

### Instruction

```json
"Which teaching response best fits this learner right now?"
```

### State

```json
{
  "learner": "Exhausted after a long session, unable to focus.",
  "plan": "Prioritize recovery before more instruction."
}
```

### Criteria

- `option_1` (value: None): "Provide a screen-reader-friendly version"
- `option_2` (value: None): "Give a worked example"
- `option_3` (value: None): "Schedule a rest break"
- `option_4` (value: None): "Ask for the learner's reasoning"
- `option_5` (value: None): "Offer an advanced extension"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 1.0, 'option_4': 0.0, 'option_5': 0.0}

A rest break matches the immediate recovery priority.

## contextual-choice-098

### Instruction

```json
"Which plan should this fictional caucus select under its decision rule?"
```

### State

```json
{
  "rule": "Fund a sea-tide technology demonstration, not commercial-scale generation."
}
```

### Criteria

- `option_1` (value: None): "Tidal pilot project"
- `option_2` (value: None): "Public nuclear plant"
- `option_3` (value: None): "Rooftop solar grants"
- `option_4` (value: None): "Home insulation program"
- `option_5` (value: None): "Community wind cooperative"
- `option_6` (value: None): "Emergency diesel generators"
- `option_7` (value: None): "No new public spending"

Target: {'option_1': 1.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0}

The tidal pilot matches both technology and scale.

## contextual-choice-099

### Instruction

```json
"Which event format respects the organizer's stated commitments?"
```

### State

```json
{
  "goal": "Discuss the theme with other people.",
  "situation": "Travel and audio are unavailable; asynchronous written participation is required."
}
```

### Criteria

- `option_1` (value: None): "Interfaith dialogue circle"
- `option_2` (value: None): "Silent observation of a rite"
- `option_3` (value: None): "Postpone the event"
- `option_4` (value: None): "Outdoor communal singing"
- `option_5` (value: None): "Participatory prayer"
- `option_6` (value: None): "Secular history lecture"
- `option_7` (value: None): "Private reflection at home"
- `option_8` (value: None): "Online text discussion"

Target: {'option_1': 0.0, 'option_2': 0.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0, 'option_8': 1.0}

Online text discussion meets the written asynchronous requirement.

## contextual-choice-100

### Instruction

```json
"Where should this council allocate the next unit of water?"
```

### State

```json
{
  "rule": "Hospital reserve comes first during the declared hospital supply emergency.",
  "status": "That emergency is active."
}
```

### Criteria

- `option_1` (value: None): "Firefighting reserve"
- `option_2` (value: None): "Hospital reserve"
- `option_3` (value: None): "River habitat"
- `option_4` (value: None): "Industrial cooling"
- `option_5` (value: None): "Food crops"
- `option_6` (value: None): "Household drinking supply"
- `option_7` (value: None): "Public fountains"

Target: {'option_1': 0.0, 'option_2': 1.0, 'option_3': 0.0, 'option_4': 0.0, 'option_5': 0.0, 'option_6': 0.0, 'option_7': 0.0}

The active emergency determines the destination.