INDEX

Explanations

HTML closing tags or tag fragments

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

empty or self-closing XML/HTML tags.

oai_token-act-pair · claude-4-5-sonnet Triggered by @bfgcn7frf5

This neuron detects the specific character pattern where an element’s start‐tag ends (quote + “>”) and is immediately followed by the start of its end‐tag (“</”), i.e. the sequence '"></'.

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

 have

-1.69

౭

-1.56

嫡

-1.48

そういえば

-1.45

our

-1.43

というわけで

-1.40

 saltar

-1.39

 pensar

-1.38

lässlich

-1.37

ссора

-1.35

POSITIVE LOGITS

1.54

in

1.41

 Comentarios

1.34

romeda

1.32

></

1.31

 able

1.30

 oyuncak

1.30

panes

1.28

 speelt

1.28

raja

1.27

Activations Density 0.001%

HTML closing tags or tag fragments

empty or self-closing XML/HTML tags.

This neuron detects the specific character pattern where an element’s start‐tag ends (quote + “>”) and is immediately followed by the start of its end‐tag (“</”), i.e. the sequence '"></'.

No Comments

No Known Activations

HTML closing tags or tag fragments

empty or self-closing XML/HTML tags.

This neuron detects the specific character pattern where an element’s start‐tag ends (quote + “>”) and is immediately followed by the start of its end‐tag (“</”), i.e. the sequence '"></'.

No Comments

No Known Activations