I don’t know what model codex uses for session summarization (I use Pro subscription, no third party models), but I get Chinese summaries from time to time, when the only Chinese that could have appeared in the session would be an i18n strings file that it may or may not have loaded. Very puzzling. Last happened yesterday.
I don’t know what model codex uses for session summarization (I use Pro subscription, no third party models), but I get Chinese summaries from time to time, when the only Chinese that could have appeared in the session would be an i18n strings file that it may or may not have loaded. Very puzzling. Last happened yesterday.