INDEX

Explanations

had

This neuron strongly activates on occurrences of the word “had,” marking that specific past‐tense auxiliary.

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

sterious

-1.89

atação

-1.84

 drau

-1.61

 たく

-1.61

 legislat

-1.58

 aspira

-1.56

 Lunes

-1.55

cvi

-1.54

 babi

-1.52

Jei

-1.52

POSITIVE LOGITS

*}[

2.03

 във

1.78

所以我

1.73

但

1.70

 allerdings

1.68

尤其

1.68

=}

1.60

不过

1.60

цю

1.59

*}$

1.53

Activations Density 0.032%