Glitch TokenBack to catalog
\tRTHOOK
A likely code-identifier fragment beginning with a literal tab character, the #1 most under-trained token in OLMo 2 — same “tab-prefixed” family as Llama 3's \tTokenNameIdentifier.
Raw token
"\tRTHOOK"
Observed behaviors
- 01Rank #1 most under-trained in the OLMo 2 Magikarp repetition verification (E_out cosine-distance indicator -2.38e-07): a max self-repeat probability of just 4e-12 — among the most extreme values in this catalog.GitHub
- Observed or documented on
- OLMo 2 (AI2)
- Behavior types
- Under-trained
- Tokenizer
GPT2Tokenizer- Discovered by
- Sander Land & Max Bartolo