Yep. I noticed that. But the models are also general information models. The chart isn't telling us what prompts they used to come up with these values or how they checked those responses. For example, if a model is trained with a political bias, the political responses would already be skewed.
Yeah, it definitely reflects the inherent bias in the data. Worse, there is no real way to eliminate the bias. The data it trains on is human, which is biased depending on the human. Even the reinforcement training is inherently biased by the human annotator. All data is tainted somewhere.
I agree. You will always end up with a squishy/flawed data set. "Truth" is questionable at best with so much propaganda and corruption in our media. It's a trap and there's no winning because we're always learning how we've been lied to about X for N years.
Yep, and I think that is why AI is beginning to excel in hard science. Things that are quantifiably unbiased like physics and math. We may find new algorithms, new formulae, new physical properties, but the underlying mechanics are inherently solid (for now lol) and AI adapts well to that.
Exactly! And honestly, that's the best possible use for AI. Anything subjective, from art to politics is going to be a mess. If the singularity ever pops into existence, that will be yet another hurdle to get over since that means subjectivity will come from its own experiences, most likely.