Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Is the reversal curse in LLMs real?

andrewmayne.com

Post reply on HN
Print view View original

131–140 of 211 posts

Re: Is the reversal curse in LLMs real?

#131

Post by RandomLensman »

dragonwriter wrote: post #68
zapperdulchen wrote: post #65

I played along the example of the chancellor question. If you deviate a little from the examples in the article GPT-4 gets it all wrong: Who is the eighth Federal Chancellor of the Federal Republic of Germany? - Olaf Scholz (wrong, Angela Merkel) https://chat.openai.com/share/937795ea-bd91-43ee-bd76-1a125f... Who was the eighth Federal Chancellor of the Federal Republic of Germany? - Helmut Kohl (wrong, Angela Merkel…

> Even when allowed to surf the web, ChatGPT gets it wrong: > Who was the eighth Federal Chancellor of the Federal Republic of Germany? - Gerhard Schröder Its probably counting Walter Scheel, who was Acting Chancellor in 1974, and you are probably not. When I asked ChatGPT to list the chancellors in order and identify the eighth, it listed eight ending in Schröder, with Scheel and his ten day Acting Chancellor tenure…

Walter Scheel was never Chancellor, he was fulfilling the duties of the Chancellor while being Foreign Minister, but he was not a Chancellor. So counting him is wrong.

Re: Is the reversal curse in LLMs real?

#132

Post by bob1029 »

user_named wrote: post #130
bob1029 wrote: post #127

Earlier quoted context omitted.

This is approximately where I wound up. I had an initial knee-jerk reaction after reading the paper, but your kind of reasoning brought me back down to earth. I don't know that I even want these things to generalize. I get it in academic terms, but from a CTO perspective how much this matters to the business? We already know our policies and procedures and aren't exactly excited about the idea of something getting cu…

Then you don't want an LLM you want something deterministic

There is no opportunity for nuance?

Re: Is the reversal curse in LLMs real?

#135

Post by Kiro »

RandomLensman wrote: post #131
dragonwriter wrote: post #68

Earlier quoted context omitted.

> Even when allowed to surf the web, ChatGPT gets it wrong: > Who was the eighth Federal Chancellor of the Federal Republic of Germany? - Gerhard Schröder Its probably counting Walter Scheel, who was Acting Chancellor in 1974, and you are probably not. When I asked ChatGPT to list the chancellors in order and identify the eighth, it listed eight ending in Schröder, with Scheel and his ten day Acting Chancellor tenure…

Walter Scheel was never Chancellor, he was fulfilling the duties of the Chancellor while being Foreign Minister, but he was not a Chancellor. So counting him is wrong.

Imagine reading this comment thread a few years back. Just getting an answer to this question at all from a generic bot would be considered science-fiction and now we're dismissing it based on technicalities.

Re: Is the reversal curse in LLMs real?

#136

Post by RandomLensman »

Kiro wrote: post #135
RandomLensman wrote: post #131

Earlier quoted context omitted.

Walter Scheel was never Chancellor, he was fulfilling the duties of the Chancellor while being Foreign Minister, but he was not a Chancellor. So counting him is wrong.

Imagine reading this comment thread a few years back. Just getting an answer to this question at all from a generic bot would be considered science-fiction and now we're dismissing it based on technicalities.

Not a technicality. It's like inventing a proper head of government that wasn't there.

Also, knowing German chancellors was always easy, so that fact that a machine falls behind any normal dictionary or the German Chancellor's website is poor form.

It's like a calculator getting basic addition wrong. I don't want a future full of poorly performing machines.

Re: Is the reversal curse in LLMs real?

#137

Post by Kiro »

RandomLensman wrote: post #136
Kiro wrote: post #135

Earlier quoted context omitted.

Imagine reading this comment thread a few years back. Just getting an answer to this question at all from a generic bot would be considered science-fiction and now we're dismissing it based on technicalities.

Not a technicality. It's like inventing a proper head of government that wasn't there. Also, knowing German chancellors was always easy, so that fact that a machine falls behind any normal dictionary or the German Chancellor's website is poor form. It's like a calculator getting basic addition wrong. I don't want a future full of poorly performing machines.

He was acting Chancellor. It's like saying a red apple is not an apple. It's a technicality. The fact that we have a completely reasonable explanation to why GPT-4 includes him in the list and you guys still make it sound like the AI is a complete moron is just hilarious to me.

Re: Is the reversal curse in LLMs real?

#138

Post by RandomLensman »

Kiro wrote: post #137
RandomLensman wrote: post #136

Earlier quoted context omitted.

Not a technicality. It's like inventing a proper head of government that wasn't there. Also, knowing German chancellors was always easy, so that fact that a machine falls behind any normal dictionary or the German Chancellor's website is poor form. It's like a calculator getting basic addition wrong. I don't want a future full of poorly performing machines.

He was acting Chancellor. It's like saying a red apple is not an apple. It's a technicality. The fact that we have a completely reasonable explanation to why GPT-4 includes him in the list and you guys still make it sound like the AI is a complete moron is just hilarious to me.

There is no "Acting Chancellor", that is a poor translation of what the job is. There is only a Chancellor, not an "Acting Chancellor" - you have to trust that the Germans and their institutions kind of know who was Chancellor and who was not.

The machine is a complete moron for not being able to get basic data from basic sources right.

Re: Is the reversal curse in LLMs real?

#139

Post by amelius »

carlmr wrote: post #111
zapperdulchen wrote: post #65

I played along the example of the chancellor question. If you deviate a little from the examples in the article GPT-4 gets it all wrong: Who is the eighth Federal Chancellor of the Federal Republic of Germany? - Olaf Scholz (wrong, Angela Merkel) https://chat.openai.com/share/937795ea-bd91-43ee-bd76-1a125f... Who was the eighth Federal Chancellor of the Federal Republic of Germany? - Helmut Kohl (wrong, Angela Merkel…

>Who is the eighth Federal Chancellor of the Federal Republic of Germany? - Olaf Scholz (wrong, Angela Merkel) Technically there's no right answer to this question. "Is" implies present. But Angela Merkel isn't the present chancellor. Olaf Scholz is. But he's not the eigth.

Language is imprecise. LLMs should be able to deal with that.

Re: Is the reversal curse in LLMs real?

#140

Post by PoignardAzur »

akoboldfrying wrote: post #9

Only skimmed this and didn't read the underlying paper, but I was surprised to see no mention of the fact that "A is B" often does not at all imply "B is A" in everyday language: A bird is an animal, but it's wrong to conclude that an arbitrary animal must be a bird.

"A is B" is meant a bit loosely here. A better shorthand would be "A is the B" "The B is A", or "A does B" "B is done by A".

The actual paper explores equivalences that actually make sense, and whether LLMs have trouble with them.

Post reply on HN