AI text detection varies by topic: SEO slop prompts almost always caught
amplifiedamp · x · 2026-11-02
The author shares a finding: detection effectiveness for AI-generated text varies mostly by topic—prompts asking the model to write SEO slop were always caught regardless of retries, with two real prompt examples shown. The augmentation is also largely length-insensitive: a higher input-to-output length ratio increases the miss rate by at most 5%.
Related event: Study Tests AI Text Detection: SEO Spam Nearly Always Caught(3 posts)→
More from Research
- Parallel Language Models: generating multiple consistent tokens per pass to cut autoregressive latency — burkov · 2026-11-02
- UnEZ paper: fully unsupervised neural network with directly interpretable latent space — hahahahaha369 · 2026-11-02
- 8 Coding Agents × 250K Context Each, Concurrently, on Two RTX 3090s (~600+ tok/s) — o2beast · 2026-11-02
- QwenBend: Qwen writes and runs Bend programs inside its own forward pass using 53M hand-set weights — sull · 2026-11-02
- Robotics needs its own CERN or Hubble to cut through the noise — DJiafei · 2026-11-02
- Hidden gems tool surfaces underrated AI papers like TinyStories and 1-bit LLMs — omarsar0 · 2026-11-02