INDEX

Explanations

identi, inver, inven

The neuron strongly activates on tokens containing the letter sequence “ent” or “end” (e.g. end, ent, identi-, inven-, inver-).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

🔭

-0.81

jusi

-0.80

Critic

-0.79

MIDDLE

-0.77

HTTPS

-0.77

Middle

-0.76

 MIDDLE

-0.76

*************

-0.73

 Middle

-0.72

RUNTIME

-0.71

POSITIVE LOGITS

 gefahren

0.82

 attemp

0.79

ところは

0.77

 ident

0.75

Yess

0.74

 Ethnicity

0.74

 gelegt

0.74

 intending

0.73

कर्ष

0.73

 indign

0.72

Activations Density 0.030%