Accurately identifies a unique target entity by precisely calculating and applying complex, multi-stage temporal constraints, including ranges and specific date components, to distinguish it from temporally similar alternatives.
Description: Accurately identifies a unique target entity by precisely calculating and applying complex, multi-stage temporal constraints, including ranges and specific date components, to distinguish it from temporally similar alternatives.
Layer 2: Guidelines
General Guidelines
Always derive a precise temporal window (start/end dates) for events, even if the query uses fuzzy language.
Ensure all components of a date (year, month, day) are considered when identifying a temporally constrained entity.
Validate the derived temporal window against all parts of the query's temporal description.
Recognize that 'give or take' implies a specific numerical range around a central value.
Specific Guidelines
For 'X years, give or take Y', calculate the range [X-Y, X+Y] relative to the base event.
When identifying an entity based on an event date, prioritize candidates that match the derived month and day, not just the year.
相關技能
If multiple entities fall within a derived year range, use specific month/day constraints to narrow down to the unique instance.
Ensure the base year for relative temporal calculations is accurately identified from its own constraints (e.g., '1980s' implies a range for the travel year).
Layer 3: Context
Problem Pattern: Incorrect identification of a target entity due to miscalculation or misapplication of complex, multi-stage temporal constraints, especially those involving ranges or specific date components, leading to the selection of a temporally similar but incorrect instance.
State Signature: Target entity identified, but its temporal attributes do not precisely match the derived temporal window from the query's complex, multi-part temporal conditions.
Bypass Strategy:
Deconstruct all temporal constraints into their base components (e.g., start year, end year, relative offsets, specific month/day).
Systematically derive the precise temporal window (year range, specific month, specific day) for the target entity's key event.
Cross-reference the derived temporal window with all potential candidate entities to identif
...