To successfully reason in the way it did, ChatGPT would have needed a meta-representation for the word âactually,â in order to understand that its prior answer was incorrect.
What makes this a meta-representation instead of something next-word-weight-y, like merely associating the appearance of âActually,â with a goal that the following words should be negatively correlated in the corpus with the words that were in the previous message?
Iâm also wondering if the butcher shop and the grocery store didnât have different answers because of the name you gave the store. Maybe it was because you gave the quantity in pounds instead of in items?
You previously told ChatGPT âThatâs because youâre basically taking (and wasting) the whole item.â ChatGPT might not have an association between âpoundâ and âitemâ the way a âcalzoneâ is an âitem,â so it might not use your earlier mention of âitemâ as something that should affect how it predicts the words that come after âpound.â
Or ChatGPT might have a really strong prior association between pounds â mass â [numbers that show up as decimals in texts about shopping] that overrode your earlier lesson.
More good points⌠I would say to refer to my reply above (which I had not yet posted when you made this comment). Just to summarize, the overall thesis stands since enough words would have needed to have meta-reps, even if we donât know particulars. Itâs easier to isolate individual words having meta-reps in the second and third sessions (I believe). In any case, thanks for helping me to drill down on this!
Thatâs a good point. We âactuallyâ canât be certain that a meta-representation was formed for a particular word in that example. I should have used the word âprobablyâ when talking about meta-representations for individual words. However, we can be fairly confident that ChatGPT formed meta-representations for enough words to go from getting an incorrect answer to a correct answer in the example. I believe we can isolate specific words better in the second and third sessions.
As far as whether associating the word âactuallyâ âwith a goal that the following words should be negatively correlated in the corpus with the words that were in the previous message,â the idea of âhaving a goalâ and the concept that the word âactuallyâ goes with negative correlations seem a bit like signs of meta-representations in and of themselves, but I guess thatâs just my opinion.
With respect to all the sessions, there may very well be similar conversations in the corpus of training data in which Person A (like myself) teaches Person B (like ChatGPT), and ChatGPT is just imitating that learning process (by giving the wrong answer first and then âlearningâ), but I address why that is probably not the case in the âmimicking learningâ paragraph.
I would say my overall thesis still stands (since enough words must have meta-reps), but good point on the particulars. Thank-you!
What makes this a meta-representation instead of something next-word-weight-y, like merely associating the appearance of âActually,â with a goal that the following words should be negatively correlated in the corpus with the words that were in the previous message?
Iâm also wondering if the butcher shop and the grocery store didnât have different answers because of the name you gave the store. Maybe it was because you gave the quantity in pounds instead of in items?
You previously told ChatGPT âThatâs because youâre basically taking (and wasting) the whole item.â ChatGPT might not have an association between âpoundâ and âitemâ the way a âcalzoneâ is an âitem,â so it might not use your earlier mention of âitemâ as something that should affect how it predicts the words that come after âpound.â
Or ChatGPT might have a really strong prior association between pounds â mass â [numbers that show up as decimals in texts about shopping] that overrode your earlier lesson.
More good points⌠I would say to refer to my reply above (which I had not yet posted when you made this comment). Just to summarize, the overall thesis stands since enough words would have needed to have meta-reps, even if we donât know particulars. Itâs easier to isolate individual words having meta-reps in the second and third sessions (I believe). In any case, thanks for helping me to drill down on this!
Thatâs a good point. We âactuallyâ canât be certain that a meta-representation was formed for a particular word in that example. I should have used the word âprobablyâ when talking about meta-representations for individual words. However, we can be fairly confident that ChatGPT formed meta-representations for enough words to go from getting an incorrect answer to a correct answer in the example. I believe we can isolate specific words better in the second and third sessions.
As far as whether associating the word âactuallyâ âwith a goal that the following words should be negatively correlated in the corpus with the words that were in the previous message,â the idea of âhaving a goalâ and the concept that the word âactuallyâ goes with negative correlations seem a bit like signs of meta-representations in and of themselves, but I guess thatâs just my opinion.
With respect to all the sessions, there may very well be similar conversations in the corpus of training data in which Person A (like myself) teaches Person B (like ChatGPT), and ChatGPT is just imitating that learning process (by giving the wrong answer first and then âlearningâ), but I address why that is probably not the case in the âmimicking learningâ paragraph.
I would say my overall thesis still stands (since enough words must have meta-reps), but good point on the particulars. Thank-you!