Answers are grounded in this paper and cite their sources.
This record archives two items connected to the TERRA preprint on SSRN (https://ssrn.com/abstract=6336958; DOI 10.2139/ssrn.6336958). ssrn-6336958.pdf: the May 2026 manuscript (v3), "Recognition-Conditioned Stability in Human–AI Interaction: The TERRA Framework", preserved as posted before the consolidated v4 revision replaces it on SSRN. terra-v4-pilot-supplement.zip: the companion archive for the consolidated v4 revision (1 October 2026). It contains the five original response CSVs (150 saved responses) from the exploratory llama3.1:8b pilot, the configuration and summary files, the original v3 runner, a portable retrospective audit script, and a relative-path SHA-256 manifest. The audit requires Python 3 and its standard library, makes no network or model calls, and reproduces the saved component scores and threshold summaries. The consolidated v4 manuscript, "TERRA and RDSP: Recognition-Conditioned Stability in Human–AI Interaction", is posted at the SSRN address above. Sampling seeds, model digest, and API metadata for the historical pilot were not recorded and are not included.
No takes yet. Share an insight, caveat, or question.
Heizer Marval-Ospino (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: