LIVEdataset aec-bench@releasetasks 552models 18last submission … · built …
electricalwith-tool

Overlap Calculation

Calculates a reduced rail signalling overlap distance from approach speed, braking rate, track gradient, reaction time, danger point distance, and low-adhesion factor. The template reports full-speed overlap, braking-only timed overlap, and residual danger point clearance.

with-tool: The model is given an executable Python calculator script.

Standards

AS 7711Network Rail NR/L2/SIG/10158

Tags

brakingdeterministicelectricaloverlaprailsignalling

How this task is generated

One template produces many comparable benchmark tasks while keeping the scoring contract fixed.

  1. 01

    Template

    The reusable contract shown on this page.

  2. 02

    Scenario

    An archetype and site context are sampled.

  3. 03

    Difficulty tier

    Inputs may be hidden at harder tiers.

  4. 04

    Task prompt

    The model responds with the declared outputs.

Parameters

Inputs the model receives, and the outputs it is scored on.

Inputs

6 inputs

Always given

Included directly in every task prompt.

5

  • Maximum approach speed kmh

    maximum_approach_speed_kmh

    Maximum train approach speed

    10 – 160 km/h

  • Emergency braking rate

    emergency_braking_rate_m_s2

    Emergency braking deceleration rate before adhesion and gradient adjustments

    0.3 – 1.5 m/s2

  • Track gradient pct

    track_gradient_pct

    Track gradient, positive for rising grade in the direction of travel

    -4 – 4 %

  • Reaction time

    reaction_time_s

    Reaction or system allowance time

    0.5 – 5 s

  • Danger point distance

    danger_point_distance_m

    Available distance from stop signal to danger point

    20 – 2000 m

Hidden at higher difficulty

Visible in easier tasks and withheld in one or more harder tiers.

1

  • Low adhesion factor

    low_adhesion_factor

    Multiplier applied to braking rate for low adhesion

    Hidden at hard difficulty.

    0.4 – 1

Scored outputs

6 outputs

Approach speed m s

approach_speed_m_s

Approach speed converted to metres per second

Scores if within ±3% of the reference value.

Gradient adjusted braking rate m s2

gradient_adjusted_braking_rate_m_s2

Effective braking rate after adhesion and gradient effects

Scores if within ±3% of the reference value.

Reaction distance

reaction_distance_m

Distance travelled during reaction time

Scores if within ±3% of the reference value.

Full speed overlap

full_speed_overlap_m

Full-speed overlap including reaction and braking distance

Scores if within ±3% of the reference value.

Timed overlap option

timed_overlap_option_m

Braking-only timed overlap distance

Scores if within ±3% of the reference value.

Danger point clearance

danger_point_clearance_m

Residual distance to the danger point after full-speed overlap

Scores if within ±3% of the reference value.

Difficulty

Each template is sampled at three tiers. Harder tiers may hide inputs, forcing the model to infer them from the scenario description.

easy

All inputs given

Metro overlap case with all values visible

medium

All inputs given

Signal overlap case selected from metro or regional lines

hard

Some inputs hidden

Regional signal with low adhesion factor embedded in context

Hidden inputs

  • Low adhesion factorlow_adhesion_factor

Prompt replacement text

Use the low adhesion factor implied by {{ archetype.description }}.

Low adhesion factor restricted to: 0.75

Task bundle

The exact instruction and parameter contract used to generate this task, pinned to the published library source.

/workspace

  • instruction.md
  • overlap-calculation_calc.py

Teal lines show Jinja input conditions, not task visibility policy. A line renders only when that input or tool is visible.

1# ABOUTME: Prompt template for rail signal overlap calculation tasks.2# ABOUTME: Presents speed, braking, gradient, reaction, adhesion, and clearance data.3 4You are a senior rail signalling engineer checking overlap distance.5 6## Given7 8| Parameter | Value | Unit |9|-----------|-------|------|10| Maximum approach speed | {{ maximum_approach_speed_kmh }} | km/h |11| Emergency braking rate | {{ emergency_braking_rate_m_s2 }} | m/s2 |12| Track gradient | {{ track_gradient_pct }} | % |13| Reaction time | {{ reaction_time_s }} | s |14| Danger point distance beyond signal | {{ danger_point_distance_m }} | m |15| Low adhesion factor | {{ low_adhesion_factor }} | - |16 17## Constraints18 19- Convert speed from km/h to m/s using division by 3.6.20- Gradient-adjusted braking rate equals emergency braking rate times low adhesion factor plus `9.81 * gradient / 100`.21- Reaction distance equals speed times reaction time.22- Timed overlap option equals `speed^2 / (2 * gradient-adjusted braking rate)`.23- Full-speed overlap equals reaction distance plus timed overlap option.24- Danger point clearance equals danger point distance minus full-speed overlap.25 26## Output Format27 28Include a JSON block with exactly these keys:29 30```json31{32 "approach_speed_m_s": <numeric_value>,33 "gradient_adjusted_braking_rate_m_s2": <numeric_value>,34 "reaction_distance_m": <numeric_value>,35 "full_speed_overlap_m": <numeric_value>,36 "timed_overlap_option_m": <numeric_value>,37 "danger_point_clearance_m": <numeric_value>38}39```40 41Write your complete solution to `/workspace/output.md`.42

Scenario archetypes

Each generated task is drawn from one of these realistic scenario bands.

Site contexts ground each scenario in a real locale the model can use to infer hidden values.

Metro signal

metro_signal

Metro rail signal overlap

metro-railsignal-overlap
Parameter ranges
maximum_approach_speed_kmh
30 – 90
emergency_braking_rate_m_s2
0.7 – 1.2
track_gradient_pct
-2.5 – 2.5
reaction_time_s
1 – 3
danger_point_distance_m
150 – 900
low_adhesion_factor
0.7 – 1

Regional signal

regional_signal

Regional rail signal overlap

regional-railsignal-overlap
Parameter ranges
maximum_approach_speed_kmh
70 – 140
emergency_braking_rate_m_s2
0.5 – 1
track_gradient_pct
-3 – 2
reaction_time_s
1.5 – 4
danger_point_distance_m
300 – 1800
low_adhesion_factor
0.6 – 0.95

Example task

regional-rail-regional-signal-preview — hard difficulty, some inputs hidden.

Regional rail signal overlap. regional-rail. Required outputs: approach_speed_m_s, gradient_adjusted_braking_rate_m_s2, reaction_distance_m, full_speed_overlap_m, timed_overlap_option_m, danger_point_clearance_m

The model sees

Scenario context and visible inputs.

maximum_approach_speed_kmh
70 to 140 km/h
emergency_braking_rate_m_s2
0.5 to 1 m/s2
track_gradient_pct
-3 to 2 %
reaction_time_s
1.5 to 4 s
danger_point_distance_m
300 to 1800 m

Executable tool: overlap-calculation_calc.py

The model must infer

Inputs withheld at this difficulty.

  • Low adhesion factor

    low_adhesion_factor

Stand-in text in the prompt

Use the low adhesion factor implied by {{ archetype.description }}.

The model must produce

The scored JSON answer schema.

{
  "approach_speed_m_s": <number>,
  "gradient_adjusted_braking_rate_m_s2": <number>,
  "reaction_distance_m": <number>,
  "full_speed_overlap_m": <number>,
  "timed_overlap_option_m": <number>,
  "danger_point_clearance_m": <number>
}
  • approach_speed_m_s · scored within ±3%
  • gradient_adjusted_braking_rate_m_s2 · scored within ±3%
  • reaction_distance_m · scored within ±3%
  • full_speed_overlap_m · scored within ±3%
  • timed_overlap_option_m · scored within ±3%
  • danger_point_clearance_m · scored within ±3%