Stochastic Parrot
A critical metaphor for large language models, originating from a 2021 paper by Emily Bender and colleagues. It characterizes these models as assembling language fragments based on statistical patterns rather than genuine semantic understanding, producing outputs that are statistically plausible yet lack awareness of word meaning, causality, or real-world grounding.
Core Facts
Timeline (last 90 days)
The "stochastic parrots" camp argues that LLMs have never been exposed to the real world that linguistic symbols refer to, and merely stitch together text that looks plausible in a statistical sense.
The stochastic parrot hypothesis, proposed by Bender et al. in 2021, questions whether large language models truly understand the meaning of language or merely reproduce statistical patterns in training data.
The concept of the Stochastic Parrot comes from the 2021 paper "On the Dangers of Stochastic Parrots" by Emily Bender, Timnit Gebru, and others.
All Facts (3)
The stochastic parrot hypothesis, proposed by Bender et al. in 2021, questions whether large language models truly understand the meaning of language or merely reproduce statistical patterns in training data.
75%UnverifiedThe concept of the Stochastic Parrot comes from the 2021 paper "On the Dangers of Stochastic Parrots" by Emily Bender, Timnit Gebru, and others.
60%UnverifiedThe "stochastic parrots" camp argues that LLMs have never been exposed to the real world that linguistic symbols refer to, and merely stitch together text that looks plausible in a statistical sense.
50%