LLMs Can Get "Brain Rot": A Pilot Study on Twitter/X

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Xing, Shuo, Hong, Junyuan, Wang, Yifan, Chen, Runjin, Zhang, Zhenyu, Grama, Ananth, Tu, Zhengzhong, Wang, Zhangyang
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908985218564096
author Xing, Shuo
Hong, Junyuan
Wang, Yifan
Chen, Runjin
Zhang, Zhenyu
Grama, Ananth
Tu, Zhengzhong
Wang, Zhangyang
author_facet Xing, Shuo
Hong, Junyuan
Wang, Yifan
Chen, Runjin
Zhang, Zhenyu
Grama, Ananth
Tu, Zhengzhong
Wang, Zhangyang
contents We propose and test the LLM Brain Rot Hypothesis: continual exposure to junk web text induces lasting cognitive decline in large language models (LLMs). To unveil junk effects, we designed a novel controlled experiment on real Twitter/X corpora, by constructing junk and reverse-controlled datasets via two orthogonal operationalizations: M1 (engagement degree) and M2 (semantic quality), with matched token scale and training operations across conditions. Compared to the control group, continual pre-training of 4 LLMs on the junk dataset causes non-trivial declines (Hedges' g>0.3) on reasoning, long-context understanding, safety, and inflating "dark traits" (e.g., psychopathy, narcissism). The gradual mixtures of junk and control datasets also yield dose-response cognition decay: for example, under M1, ARC-Challenge with Chain-of-Thought drops 72.1 -> 57.2 and RULER-CWE 83.7 -> 52.3 as junk ratio rises from 0% to 100%. Error forensics reveal several key insights. First, we identify thought-skipping as the primary lesion in reasoning: models increasingly truncate or skip chains. Second, partial but incomplete healing is observed: scaling instruction tuning and clean continual pre-training improve the declined cognition, yet cannot restore baseline capability, suggesting persistent representational drift rather than format mismatch. Finally, we discover that the popularity, a non-semantic metric, of a tweet is a better indicator of the Brain Rot effect than the length in M1. Together, the results provide significant, multi-perspective evidence that social effects of data could be a causal driver of LLM capability decay in continual pre-training, thereby motivating routine "cognitive health checks" for deployed and evolving LLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2510_13928
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LLMs Can Get "Brain Rot": A Pilot Study on Twitter/X
Xing, Shuo
Hong, Junyuan
Wang, Yifan
Chen, Runjin
Zhang, Zhenyu
Grama, Ananth
Tu, Zhengzhong
Wang, Zhangyang
Computation and Language
Artificial Intelligence
We propose and test the LLM Brain Rot Hypothesis: continual exposure to junk web text induces lasting cognitive decline in large language models (LLMs). To unveil junk effects, we designed a novel controlled experiment on real Twitter/X corpora, by constructing junk and reverse-controlled datasets via two orthogonal operationalizations: M1 (engagement degree) and M2 (semantic quality), with matched token scale and training operations across conditions. Compared to the control group, continual pre-training of 4 LLMs on the junk dataset causes non-trivial declines (Hedges' g>0.3) on reasoning, long-context understanding, safety, and inflating "dark traits" (e.g., psychopathy, narcissism). The gradual mixtures of junk and control datasets also yield dose-response cognition decay: for example, under M1, ARC-Challenge with Chain-of-Thought drops 72.1 -> 57.2 and RULER-CWE 83.7 -> 52.3 as junk ratio rises from 0% to 100%. Error forensics reveal several key insights. First, we identify thought-skipping as the primary lesion in reasoning: models increasingly truncate or skip chains. Second, partial but incomplete healing is observed: scaling instruction tuning and clean continual pre-training improve the declined cognition, yet cannot restore baseline capability, suggesting persistent representational drift rather than format mismatch. Finally, we discover that the popularity, a non-semantic metric, of a tweet is a better indicator of the Brain Rot effect than the length in M1. Together, the results provide significant, multi-perspective evidence that social effects of data could be a causal driver of LLM capability decay in continual pre-training, thereby motivating routine "cognitive health checks" for deployed and evolving LLMs.
title LLMs Can Get "Brain Rot": A Pilot Study on Twitter/X
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2510.13928