3 ms·
Of course, the initial mistake is assuming LLMs have any 'reasoning'...
by nottorp 2y ago
Of course, the initial mistake is assuming LLMs have any 'reasoning'...
- ticklemyelmo 2y ago"The breakdown is dramatic, as models also express strong overconfidence in their wrong solutions, while providing often non-sensical "reasoning"-like explanations akin to confabulations to justify and backup the validity of their clearly failed responses, making them sound plausible." It's fascinating how much they anthropomorphize the systems and credit them with emotional, possibly deceitful behaviour, in a paper trying to explain how unintelligent they are.