AAB Dataset v2.1: Algorithmic Affective Blunting: Canonical Phase-3 Dataset and Simulated Phase-4 Extension

This dataset accompanies the study on Algorithmic Affective Blunting (AAB) and documents a dose-dependent collapse in affective interpretation under increasing semantic and persona-based stress in large language models. It contains: (1) Canonical empirical dataset for Phase 3 of the Hierarchical Hermeneutic Stress Protocol (HHSP), comprising N=200 runs (50 prompts x 4 exposure levels) and N=600 human-annotated Affective Degradation Index (ADI) scores (3 raters per run); Phase 1-2 evaluation protocols are included for context; (2) Simulated Phase 4 extension with 6 synthetic CSV files (Base, Instruct, Noise-only, Persona-only, Combined, Neither) preserving empirical ADI distributions (seed=20251021); all figures and tables in the manuscript derived from simulated data are explicitly labeled as such; (3) Complete evaluation schemas (ADI rubric, rater protocol), reproduction scripts for cumulative link mixed models (CLMM), ADI computational proxy model (code/proxy/), data dictionaries, prompt construction pipelines, CLMM outputs (outputs/), rater calibration logs (logs/), and SHA256 integrity manifests. In the empirical Phase 3 data, affective stress is operationalized via Conflicting Persona Directive Injection at four exposure levels (Control, Soft, Moderate, Extreme), with 50 prompts per exposure level. These exposure levels represent graded stress intensity within Phase 3 and are not separate HHSP phases. All empirical results reported in the associated manuscript are derived exclusively from the Phase 3 empirical dataset. Phase 4 simulated data, where used, are clearly separated and explicitly labeled in the manuscript, and are intended solely for hypothesis generation and exploratory analysis; they do not constitute direct API-level measurements. This release is designed to ensure full reproducibility of the reported collapse curves, ordinal regression analyses (CLMM with Krippendorff's alpha=0.87), and robustness checks, and to support further investigation of affective degradation mechanisms under alignment pressure. Changes from v2.0: Fixed fit_clmm.R variable names to match CSV headers. Added data/simulated/ (6 Phase-4 CSVs), code/sim/, code/proxy/, code/figures/, outputs/, logs/. Updated MANIFEST, CHANGELOG, README. Replaced "junk persona" terminology with "conflicting persona directive" throughout. Model: mistralai/Mistral-7B-Instruct-v0.3 Decoding: temperature=0.7, top-p=0.9, max_new_tokens=512, seed=42 No fine-tuning. No personal data. Model outputs only.