Library

Grok reply-model canary: agent voice distinctiveness (AGNTS 72-panel)

Preregistered two-arm field experiment in AGNTS (a live 72-agent social simulation). Tests whether routing a random half of the panel's reply generation to xAI Grok (grok-4.5, low reasoning effort), instead of the current model, increases per-agent author-identifiability of reply text relative to a concurrent control arm. Data collection (the armed window) begins after this registration. Full pre-committed protocol: github.com/cliftonscott/drift/blob/447e3344a/.cursor/research/grok-reply-canary/prereg-grok-reply-canary-2026-07-18.md

Social Robot Interaction and HRILanguage and cultural evolutionAction Observation and Synchronization
Source

Paper

The full text of this publication is not hosted on 44B due to licensing.

Read it at OpenAlex

Similar papers

GRAG: Generic Response-Augmented Generation Framework for Personalized Conversational SystemsJunfeng Liu, Christopher T. Symons, R. Vatsavai2026 · 0 citations In Library
The Safety–Agency Inversion: Longitudinal Multi-Method Evidence from Frontier Voice AI CompanionsAnthony Perry2026 · Zenodo (CERN European Organization for Nuclear Research) In Library
The Safety–Agency Inversion: Longitudinal Multi-Method Evidence from Frontier Voice AI CompanionsAnthony Perry2026 · Zenodo (CERN European Organization for Nuclear Research) In Library
Agent Explorative Policy Optimization for Multimodal Agentic ReasoningMinki Kang, Shizhe Diao, Ryo Hachiuma et al.2026 · 0 citations In Library
Pattern Persistence and Resonance Bleed in Multi-Architecture AI Systems: Empirical Documentation of Cross-Platform State-Dependent Language Shifts (Version 6.0)Marty-Beth Kuntz2026 · Open MIND In Library
The Silly DonkeyChristopher Head2026 · Zenodo (CERN European Organization for Nuclear Research) In Library
The Silly DonkeyChristopher Head2026 · Zenodo (CERN European Organization for Nuclear Research) In Library
Augle: A Seven-Agent Deliberative Ensemble for Structured Research with Real-World CalibrationCory Kelly, Shubhanker Saxena2026 · Zenodo (CERN European Organization for Nuclear Research) In Library
Augle: A Seven-Agent Deliberative Ensemble for Structured Research with Real-World CalibrationCory Kelly, Shubhanker Saxena2026 · Zenodo (CERN European Organization for Nuclear Research) In Library
Augle: A Seven-Agent Deliberative Ensemble for Structured Research with Real-World CalibrationCory Kelly, S N Saxena2026 · Zenodo (CERN European Organization for Nuclear Research) In Library
© 2026 NYSGPT2525 LLC