Remix clone Hacker News

just to confirm I read this right, "the marathon crater" does not in fact exist, but this works because it seems like it should?

▲

alickz 21 hours ago | parent | next [-]

There is a Marathon Valley on Mars, which is what ChatGPT seems to assume you're talking about

https://chatgpt.com/share/680a98af-c550-8008-9c35-33954c5eac...

>Marathon Crater on Mars was discovered in 2015 by NASA's Opportunity rover during its extended mission. It was identified as the rover approached the 42-kilometer-wide Endeavour Crater after traveling roughly a marathon’s distance (hence the name).

>>is it a crater?

>>>Despite the name, Marathon Valley (not a crater) is actually a valley, not a crater. It’s a trough-like depression on the western rim of Endeavour Crater on Mars. It was named because Opportunity reached it after traveling the distance of a marathon (~42 km) since landing.

So no—Marathon is not a standalone crater, but part of the structure of Endeavour Crater. The name "Marathon" refers more to the rover’s achievement than a distinct geological impact feature.

▲

alickz 20 hours ago | parent [-]

Here's me testing with a place that is a lot less ambiguous

https://chatgpt.com/share/680aa212-8cac-8008-b218-4855ffaa20...

	▲	zapperdulchen 13 hours ago \| parent [-]
		That reaction is very different from the Marathon crater one though it uses the same pattern. I think OP's reasoning that there is a naive commitment bias doesn't hold. But to see almost all LLMs to fall into the ambiguity trap, is important for any real world use.

▲

thatjoeoverthr 21 hours ago | parent | prev | next [-]

The other aspect is it can’t reliably tell whether it „knows” something or not. It’s conditioned to imitate the corpus, but the corpus in a way is its „universe” and it can’t see the boundaries. Everything must map to something _in_ the corpus.

	▲	brookst 6 hours ago \| parent [-]
		This isn’t true — LLMs can generalize and synthesize information not in the corpus. You can ask one to create a new written language and get a grammar and vocabulary that is nowhere in the corpus.

▲

thatjoeoverthr 21 hours ago | parent | prev [-]

Yes, and the forward-only inference strategy. It seems like a normal question, so it starts answering, then carries on from there.

	▲	genewitch 6 hours ago \| parent [-]
		How come there is not the equivalent of a stable diffusion "sampler * selection for LLMs? The restart sampler for stable diffusion is so good compared to most of the original samplers. I often try to get an answer from the LM studio people but I think I really should ask the llama.cpp people.