TrajectoryResendScenario
Agents scenarios · Implementation: mlsysim/engine/scenario_definitions/agents.py.
Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.
A trajectory of turns calls that each resend the whole history (TrajectoryResendModel).
Notes
Equation: Compose mlsysim.solvers.TrajectoryResendModel.
Assumptions: The stated fields are inputs to mlsysim.solvers.TrajectoryResendModel; its documented validity regime, units, and modeling choices apply unchanged.
Excludes: This scenario adds no measurement, empirical calibration, or mechanisms beyond mlsysim.solvers.TrajectoryResendModel.
References
Derivation: this scenario binds its input fields to the referenced implementation contracts: mlsysim.solvers.TrajectoryResendModel. The methods below identify any additional arithmetic.
Examples
>>> from mlsysim import Scenarios
>>> case = Scenarios.CodingAgentResend30Turns
>>> type(case.result).__name__
'TrajectoryResendResult'Attributes
| Name | Description |
|---|---|
| appended | |
| first_prompt | |
| floors | |
| output_per_turn | |
| result | Evaluate this scenario through mlsysim.solvers.TrajectoryResendModel.solve. |
| turns |
floors
Type: DecodeScenario.
Declared default: required.
first_prompt
Type: int.
Declared default: required.
appended
Type: int.
Declared default: required.
output_per_turn
Type: int.
Declared default: required.
turns
Type: int.
Declared default: required.
result
engine.scenario_types.TrajectoryResendScenario.result
Evaluate this scenario through mlsysim.solvers.TrajectoryResendModel.solve. Equation: pass the stated fields to that implementation; its contract defines the calculation and feasibility tests. Units: each returned field follows the referenced result schema; no independent equation is evaluated by this wrapper.