LLMemotions

Changes from v1.0.0: Added reanalysis/ folder with supplementary analysis script (reanalysis.py) and outputs (valid dialogue subset comparisons, data audit, direction consistency check) Updated README with known limitations, corrected file names (event_log_365 filenames), and accurate repository structure Fixed CITATION.cff DOI placeholder Documented data exclusion criteria: ~46-49% empty LLM outputs and NaN-labeled dialogue rows are now explicitly reported The simulation code and event log data are unchanged from v1.0.0. All newly reported limitations (keyword-based state updates, combat aggregation mechanic, empty generation rate) were present in the original simulation and are now transparently documented in the manuscript and README.