LLMs Can Get "Brain Rot"!

ICLR 2026 Conference Submission18667 Authors

Keywords: LLM, safety, cognitive health, social media

TL;DR: Our work is the first to provide controlled experimental evidence that continual exposure to junk web text causally degrades LLM reasoning, instruction-following, and even personality traits.

Abstract:

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 causally isolate data quality, we run controlled experiments on real Twitter/X corpora, constructing junk and reversely controlled datasets via two orthogonal operationalizations: **M1** (engagement degree) and **M2** (semantic quality), with matched token scale and training operations across conditions. Contrary 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 Thoughts drops $74.9 \rightarrow 57.2$ and RULER-CWE $84.4 \rightarrow 52.3$ as junk ratio rises from $0$ % to $100$ %.

Error forensics reveal several key insights. First, we identify *thought-skipping as the primary lesion*: models increasingly truncate or skip reasoning chains, explaining most of the error growth. Second, partial but incomplete healing is observed: scaling instruction tuning and clean data 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 *data quality is a causal driver of LLM capability decay*, reframing curation for continual pretraining as a *training-time safety* problem and motivating routine "cognitive health checks" for deployed LLMs.

Primary Area: alignment, fairness, safety, privacy, and societal considerations

Submission Number: 18667