Cable Forum - View Single Post

**Damien** · Yesterday, 23:26

He does intervene when he doesn't like some of the answers. None of us truly knows how much interference there is on these LLMs, but Grok is the most opinionated, whereas ChatGPT and Gemini are more cautious about expressing an opinion.

Also @Grok seems to behave differently to Grok itself.

I've really been liking Gemini recently. It's quite capable, useful and good for code. I think the important thing with these is to understand how they source their knowledge and avoid more abstract or esoteric questions that require a different way of thinking.

---------- Post added at 22:26 ---------- Previous post was at 22:19 ----------

Article with more examples; https://www.theverge.com/ai-artifici...-getting-weird

I assume he'll probably have to readjust it a bit.

Yesterday, 23:26	#31
Damien Remoaner Cable Forum Team Join Date: Mar 2004 Posts: 32,904	Re: General AI discussion He does intervene when he doesn't like some of the answers. None of us truly knows how much interference there is on these LLMs, but Grok is the most opinionated, whereas ChatGPT and Gemini are more cautious about expressing an opinion. Also @Grok seems to behave differently to Grok itself. I've really been liking Gemini recently. It's quite capable, useful and good for code. I think the important thing with these is to understand how they source their knowledge and avoid more abstract or esoteric questions that require a different way of thinking. ---------- Post added at 22:26 ---------- Previous post was at 22:19 ---------- Article with more examples; https://www.theverge.com/ai-artifici...-getting-weird I assume he'll probably have to readjust it a bit. Last edited by Damien; Yesterday at 23:22.