Large language model
An AI system trained on enormous amounts of text that generates fluent language by predicting likely next tokens; abbreviated LLM.
As defined in et_941
Other definitions
A neural network trained on vast amounts of text data to predict and generate human language; examples include ChatGPT, designed to continue text in ways that match human expectations
CHATGPT won't ever be the same again after thisA neural network trained on large text corpora that generates and reasons over natural language, such as OpenAI's models, Anthropic's Claude, or Meta's LLaMA.
Dev Ittycheria - The Database Evolution - [Invest Like the Best, EP.373]
Where it comes up · 43
- Ray Kurzweil: Does AI Make Immortality Possible?11 claims
- The Future of Knowledge Graphs in a World of Large Language Models10 claims
- Geoff Hinton - Will Digital Intelligence Replace Biological Intelligence? | Vector's Remarkable 20247 claims
- The Future of Artificial Intelligence7 claims
- 343 | Tom Griffiths on The Laws of Thought7 claims
- Erik Hoel on the Threat to Humanity from AI6 claims
- 324 - Debating The Future Of Ai6 claims
- Is AI's "intelligence" an illusion? | GZERO World with Ian Bremmer6 claims
and 35 more
Most important claims · 10 of 140Search all claims for “Large language model” →
- Combining large language models with fact-checking systems (like Google or Wikipedia searches) is theoretically appealing but practically difficult because it requires translating the LLM's natural language output into formal logic, a problem AI researchers have been struggling with for 75 years without complete success.Is AI's "intelligence" an illusion? | GZERO World with Ian Bremmer
- The transformer, the core algorithm of large language models, was invented at Google and published in 2017, but Google did not productize it citing safety, so OpenAI picked it up and ran with it, and Anthropic and other US labs also build on work they did not create.Marc Andreessen - The Battle For Tech Supremacy - [Invest Like the Best, EP.410]
- An agent-based model using LLMs with survey-derived personas on a minimal social media platform—with no engagement algorithm—spontaneously produced all three pathologies of real social media: echo chambers (Democrats and Republicans not following each other), power-law attention inequality, and the 'social media prism' where more extreme/polarized users gain more attention.330 | Petter Törnberg on the Dynamics of (Mis)Information
- Because resharing/reposting is highly emotional and reactive (people share what upsets or emotionally moves them), there is a feedback loop between what gets shared and whom people follow, so emotional sharing doesn't just shape visible content but gradually builds the network structure itself—giving more polarized users more followers and attention.330 | Petter Törnberg on the Dynamics of (Mis)Information
- AI chatbots can reduce conspiracy belief: participants write their conspiracy in their own words, the AI counters the specific reasons with detailed evidence over an ~8-minute conversation, producing roughly a 20% drop in confidence and about 25% of believers no longer believing—an effect that does not decay one to two months later, which is exceptionally durable by psychology standards.333 | Gordon Pennycook on Unthinkingness, Conspiracies, and What to Do About Them
- Hallucinations in large language models are not evidence of lack of understanding—humans hallucinate constantly through constructive memory recall: human memory does not retrieve stored facts but reconstructs plausible narratives, making details often false while preserving overall validity, as demonstrated by John Dean's Watergate testimony which had false details but accurately conveyed the White House situation.Geoff Hinton - Will Digital Intelligence Replace Biological Intelligence? | Vector's Remarkable 2024
- In experiments that selectively disabled features of the debunking AI, a blunt non-persuasive fact-delivery had essentially the same effect as the original, but an AI that tried to persuade without using any facts had no effect—demonstrating that facts and evidence, not flattery or politeness, are what change conspiracy beliefs.333 | Gordon Pennycook on Unthinkingness, Conspiracies, and What to Do About Them
- Large language models illustrate exactly how much data is needed to solve Chomsky's learning problem: if LLMs have learned something like human language, this proves Chomsky's point that learning language requires enormous data—a child cannot figure out language structure in five years, and it takes something like 5,000 to 50,000 years of data to do it well; the difference is made up by inductive bias from evolution and broader world experience.343 | Tom Griffiths on The Laws of Thought
- The big difference between human minds and current large language models is inductive bias: a human child learns language from about five years of exposure, whereas the data used to train an LLM is equivalent to between 5,000 and 50,000 years of continuous speech; the thing that makes up that orders-of-magnitude gap is the inductive bias—prior distributions broadly construed—that humans bring.343 | Tom Griffiths on The Laws of Thought
- Raw language models like GPT-4 are very sensitive to the probability of the outputs they must produce: they are better at counting strings of 30 letters than 29 because the number 30 appears more often on the internet than 29, so when nearby high-probability answers exist the model produces the high-probability one rather than the correct one—an idiosyncratic bias the authors called 'embers of autoregression.'343 | Tom Griffiths on The Laws of Thought