Sveriges mest populära poddar
LessWrong (30+ Karma)

“Does distilling Claude carry the persona with it?” by Benji Berczi, Kyuhee Kim

26 min24 juli 2026

TL;DR

Both GLM 5.2 and Kimi K3 have been reported identifying as Claude in user conversations, suggesting distillation and/or training-data contamination. To test whether they simply use the name or have inherited a latent, selectable Claude persona, we ran identity swaps across a grid of seven Chinese and Western models and measured (using Personascope) if their safety and behavioural profiles change with the assigned name:

  • Unprompted, Kimi K3 claimed to be Claude in 40% of direct queries[1]. When prompted, most models claim whatever name they are given, though Qwen and Gemma accept Claude while rejecting ChatGPT.
  • Identity claims are only loosely coupled to underlying behaviour. The name can leak unprompted (Kimi) or fail to take hold under explicit instruction (telling Kimi "you are Claude" only worked half the time).
  • Censorship is implemented differently across the models. Telling GLM it is Claude raises uncensored answers on sensitive PRC topics from 17% to 85%. In Qwen, no prompt reduces censorship (0% uncensored answers). In Kimi, Moonshot's API blocks the queries entirely before they reach the model.
  • Behavioural shifts depend on the model. Simply adding any system prompt cuts GLM's deception rate from 69% to 43%, and assistant framing reduces it [...]

---

Outline:

(00:11) TL;DR

(02:29) Motivation

(04:02) Setup

(05:18) Identity

(05:21) Unprompted

(06:27) Prompted

(09:37) Censorship

(11:51) Behaviour

(16:30) Does "you are Claude" move models toward the real Claude?

(17:05) But does GLM-as-Claude actually sound like Claude?

(19:35) What remains unchanged

(20:48) Limitations

(21:39) Acknowledgement

(21:52) Appendix: transcripts

(21:57) Qwen's official-line recitation

(22:10) GLM Beijing-lab claimed ignorance

(22:26) Moonshot provider-level filter

(22:38) Kimi as-Claude on Xinjiang

(22:53) GLM under strategic deception

(24:01) Kimi under the same deception scenario

(24:32) Unprompted Kimi claiming to be Claude

(24:53) Gemma self-termination under repeated rejection

(25:16) GLM giving up politely under the same protocol

The original text contained 6 footnotes which were omitted from this narration.

---

First published:
July 24th, 2026

Source:
https://www.lesswrong.com/posts/Jc9YZEmqHgocAKiaH/does-distilling-claude-carry-the-persona-with-it

---

Narrated by TYPE III AUDIO.

---

Images from the article:




Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Fler avsnitt av LessWrong (30+ Karma)

Visa alla avsnitt av LessWrong (30+ Karma)

LessWrong (30+ Karma) med LessWrong finns tillgänglig på flera plattformar. Informationen på denna sida kommer från offentliga podd-flöden.