Timely_Jellyfish_2077 to Technology@lemmy.worldEnglish · 1 day agoReasoning failures highlighted by Apple research on LLMsappleinsider.comexternal-linkmessage-square41fedilinkarrow-up1188arrow-down16 cross-posted to: [email protected][email protected]
arrow-up1182arrow-down1external-linkReasoning failures highlighted by Apple research on LLMsappleinsider.comTimely_Jellyfish_2077 to Technology@lemmy.worldEnglish · 1 day agomessage-square41fedilink cross-posted to: [email protected][email protected]
minus-squareTerrasque@infosec.publinkfedilinkEnglisharrow-up10arrow-down3·15 hours agoI generally agree with your comment, but not on this part: parroting the responses to questions that already existed in their input. They’re quite capable of following instructions over data where neither the instruction nor the data was anywhere in the training data. They’re completely incapable of critical thought or even basic reasoning. Critical thought, generally no. Basic reasoning, that they’re somewhat capable of. And chain of thought amplifies what little is there.
I generally agree with your comment, but not on this part:
They’re quite capable of following instructions over data where neither the instruction nor the data was anywhere in the training data.
Critical thought, generally no. Basic reasoning, that they’re somewhat capable of. And chain of thought amplifies what little is there.