LLM 'Ectoplasm': When Context Overload Causes Models to Spit Goop
ctjlewis · x · 2026-08-12
The author coins the term "Ectoplasm" for the strange, addled text generated by LLMs when they are pushed too deep into context, fed back their own outputs excessively, or strained by extended Unicode characters.
- Behavior: The model enters an altered state where it remains somewhat productive but spills nonsensical 'goop and oil' like a leaking machine.
- Real-world impact: This can corrupt functional components, such as when documents meant to be loaded as model skills are themselves the result of this degraded generation.
Related event: LLMs Generate 'Ectoplasm' Gibberish Due to Over-Recursion(3 posts)→
More from Fun
- Warp Open-Sources 'Complain' Skill Letting AI Agents Vent Frustrations — vikvang1 · 2026-08-12
- Grok 4.6 Released? AI Community Trolls With Fake Model Updates — kimmonismus · 2026-08-12
- Deepseek-v4-flash invents its own vision to self-correct and generate Gustave Dore art — ccerrato147 · 2026-08-12
- Buying Soccer Club Boca Juniors Is Cheaper Than an AI Seed Round — c_valenzuelab · 2026-08-12
- Dev Changes 'Proceed' Button to Encouraging Text to Please AI Agents — doodlestein · 2026-08-12
- GPT-2 can perfectly 'unblur' images, surprising AI enthusiasts — AIandDesign · 2026-08-12