INDEX

Explanations

sequences resembling DNA or gene notation

np_acts-logits-general · gemini-2.5-flash-lite

technical alphanumeric codes, particularly those starting with 'h' such as file formats, model numbers, and technical identifiers.

oai_token-act-pair · claude-3-7-sonnet-20250219 Triggered by @neilrathi

The neuron strongly activates on tokens representing the letter “h” (upper- or lower-case) or phoneme symbols/digraphs that include “h” (e.g. ch, sh, th, jh, dh, zh, hh).

oai_token-act-pair · o4-mini Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_10/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 inspiradoras

-2.08

 because

-1.93

 consideradas

-1.91

 motivadoras

-1.90

 considerados

-1.88

 Krieges

-1.83

-1.81

 even

-1.80

you

-1.75

 encontrados

-1.73

POSITIVE LOGITS

derdag

1.72

1.67

 horloge

1.65

/)

1.63

+),

1.59

ouſly

1.56

 frau

1.55

釹

1.50

 davanti

1.49

也有

1.47

Activations Density 0.077%

sequences resembling DNA or gene notation

technical alphanumeric codes, particularly those starting with 'h' such as file formats, model numbers, and technical identifiers.

The neuron strongly activates on tokens representing the letter “h” (upper- or lower-case) or phoneme symbols/digraphs that include “h” (e.g. ch, sh, th, jh, dh, zh, hh).

No Comments

No Known Activations

sequences resembling DNA or gene notation

technical alphanumeric codes, particularly those starting with 'h' such as file formats, model numbers, and technical identifiers.

The neuron strongly activates on tokens representing the letter “h” (upper- or lower-case) or phoneme symbols/digraphs that include “h” (e.g. ch, sh, th, jh, dh, zh, hh).

No Comments

No Known Activations