INDEX

Explanations

underscore character

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

The neuron detects typographic markup tokens indicating emphasis (e.g. “Bold,” “italicized,” “emphasis”).

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

the word "Bold" at the beginning of titles or emphatic statements.

oai_token-act-pair · claude-4-5-sonnet Triggered by @jyhe0408

mentions of typographic emphasis or text formatting styles indicating emphasis.

oai_token-act-pair · gpt-5 Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ประเทศ

0.57

اری

0.57

위

0.55

Shr

0.55

Roth

0.55

위

0.55

barkeit

0.54

 kraju

0.54

Aryan

0.53

事人

0.52

POSITIVE LOGITS

 bold

1.33

 Bold

1.26

Bold

1.20

bold

1.13

BOLD

1.05

 bolder

1.05

 italic

1.03

 underlined

0.97

強調

0.97

 audacious

0.96

Activations Density 0.056%

underscore character

The neuron detects typographic markup tokens indicating emphasis (e.g. “Bold,” “italicized,” “emphasis”).

the word "Bold" at the beginning of titles or emphatic statements.

mentions of typographic emphasis or text formatting styles indicating emphasis.

No Comments

No Known Activations

underscore character

The neuron detects typographic markup tokens indicating emphasis (e.g. “Bold,” “italicized,” “emphasis”).

the word "Bold" at the beginning of titles or emphatic statements.

mentions of typographic emphasis or text formatting styles indicating emphasis.

No Comments

No Known Activations