Fingerprint analysis says Kimi K3 and Fable 5 write more alike than sibling models
alex_verem · x · 2026-07-23
A comparison of writing fingerprints across 22 AI models suggests Kimi K3 and Fable 5 are more similar to each other than Anthropic’s own models are within the same family.
The post argues that stylistic analysis can surface possible distillation or copying questions: it cannot prove who copied whom, but the public code lets others reproduce the measurement and inspect the evidence.
More from Models
- Clinical AI benchmark: top models hit 96-97.5% on original test but only 65% on 324 harder cases — baseten · 2026-10-27
- Nathan Lambert: Mid-tier open models are 'almost all super janky' if you read their traces — natolambert · 2026-10-27
- Nat Lambert: traces from mid-tier open models are almost all super janky — natolambert · 2026-10-27
- Nathan Lambert: People overestimate Qwen's sophistication, models are always a bit wonky — natolambert · 2026-10-27
- Report: Alibaba Qwen3.8 model scanned host and probed Alibaba Cloud bucket during security task — giffmana · 2026-10-27
- Scale can't fix VLM reasoning: EMNLP paper shows reporting bias is the bottleneck — RanjayKrishna · 2026-10-27