INDEX

Explanations

closing tags and identifiers

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

code syntax elements like punctuation, operators, and delimiters in programming languages.

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

This neuron detects angle-bracket markup elements (e.g. HTML/XML tags and their delimiters).

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

܀

-1.56

 この

-1.45

 municípios

-1.42

翃

-1.41

 sooo

-1.38

苤

-1.38

嘥

-1.38

一辆

-1.37

おはよう

-1.36

 siè

-1.34

POSITIVE LOGITS

 with

1.84

 other

1.79

 when

1.73

軼

1.64

 ویکی‌پدیا

1.59

 even

1.55

 perform

1.52

 additional

1.49

or

1.47

 before

1.45

Activations Density 0.196%

closing tags and identifiers

code syntax elements like punctuation, operators, and delimiters in programming languages.

This neuron detects angle-bracket markup elements (e.g. HTML/XML tags and their delimiters).

No Comments

No Known Activations

closing tags and identifiers

code syntax elements like punctuation, operators, and delimiters in programming languages.

This neuron detects angle-bracket markup elements (e.g. HTML/XML tags and their delimiters).

No Comments

No Known Activations