In my opinion still the most egregious example in history of a commercial LLM going off the rails in production. Never any technical postmortem from Google on this.
If I remember correctly, it was in fact possible to manually inject chat context at the time, which would have made spoofing something like this completely possible.
But the silence on it is very frustrating.
Point is, if Gemini is flawed then there's a very good chance that it's still deeply flawed today, and getting smarter at the same time - that is a very bad combination.
Training a new base model from scratch happens every so often. Closed labs do not publish which models are new base models but as a rule of thumb major release numbers are an indication (with some exceptions).
* - as in, Skinner psychology. The set of observable behaviors. Not speaking directly here to anything like an inner life of models.