Humans versus Machines: The Hallucination Edition
We recorded 7 claims from this post. 3 were checked against evidence. 4 can’t be checked. 2 more are held. Read the post at its source.
Claims from this post (7)
In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.
Assessment: SupportedSaid
Causal claim
“The very process that LLMs use to generalize also creates hallucinations.”
Our reading: The same process LLMs use to generalize is what creates their hallucinations.
No horizon Evidence as of
Limitation: The theory covers rare, arbitrary facts most clearly. Other kinds of hallucination may have other causes.
Assessment: MixedSaid
Observation
“Instead, everything that LLMs say -- true or false -- comes from the same process of statistically reconstructing what words are likely in some context. They NEVER fact-check what they say.”
Our reading: All LLM output, true or false, comes from the same statistical word-reconstruction process, and LLMs never fact-check what they say.
No horizon Evidence as of
Limitation: The packet documents next-token generation only for GPT-4 and does not cover every LLM in use during the period.
Assessment: Can’t be checkedSaid
Causal claim
“But even with perfect data, the stochastic reconstructive process would still produce some errors.”
Our reading: Even with perfect data, the stochastic reconstructive process LLMs use would still produce some errors.
No horizon
Qualifier: even with perfect data
Assessment: MixedSaid
Causal claim
“Instead, LLM "hallucinations" arise, regularly, because (a) they literally don't know the difference between truth and falsehood, (b) they don't have reliably reasoning processes to guarantee that their inferences are correct and (c) they are incapable of fact-checking their own work.”
Our reading: LLM hallucinations arise regularly because LLMs literally don't know the difference between truth and falsehood, don't have reasoning processes reliable enough to guarantee their inferences are correct, and are incapable of fact-checking their own work.
No horizon Evidence as of
Limitation: The verdict separates supported reliability limitations from contradicted categorical incapacity; it does not validate the full causal account.
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Rather, the goal of AGI should be to build machines that can reliably reason and plan about a wide swathe of the world.”
Our reading: The goal of AGI should be machines that reliably reason and plan about a wide swathe of the world, not recreating humans.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“The field of AI will eventually do better, but probably with an AI that is structured differently, in which facts are first-class citizens, rather than something you hope you might get for free with enough data.”
Our reading: AI will eventually do better on hallucinations, probably via differently structured systems that treat facts as first-class rather than via more data.
No horizon
Qualifier: eventually
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“TL;DR: Don't console yourself with making something that superficially looks like human errors, if you aspire to AGI.”
Our reading: Those aiming for AGI should not take comfort in LLM errors superficially resembling human errors.
No horizon
Qualifier: if you aspire to AGI
Held from this post (2)
These records did not pass every check, so nothing about them is published. Held is not a finding.
- The criteria are ambiguous. Our reading, not assessed: LLM errors result from a different underlying process than human errors: because LLMs lack intentions, egos, and financial interests, they do not lie, and they also do not take drugs and do not have emotional states.
- The two reviews read the claim differently