INDEX

Explanations

brain anomalies

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

document and citation formatting elements, particularly in academic and technical publications.

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

The neuron fires strongly on document section headers or titles—i.e. lines of consecutive capitalized words that mark an overview, background, or other named sections.

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

 layered

-1.48

 such

-1.41

—

-1.41

)$-

-1.41

 canta

-1.34

轳

-1.31

к

-1.29

菒

-1.28

 kolejne

-1.27

~”

-1.25

POSITIVE LOGITS

 they

1.82

 bereiken

1.48

 Američ

1.47

もありました

1.45

 verwijzen

1.36

 veranderen

1.34

 leveren

1.34

 pliable

1.33

 übrigens

1.31

সূত্র

1.30

Activations Density 0.067%

brain anomalies

document and citation formatting elements, particularly in academic and technical publications.

The neuron fires strongly on document section headers or titles—i.e. lines of consecutive capitalized words that mark an overview, background, or other named sections.

No Comments

No Known Activations

brain anomalies

document and citation formatting elements, particularly in academic and technical publications.

The neuron fires strongly on document section headers or titles—i.e. lines of consecutive capitalized words that mark an overview, background, or other named sections.

No Comments

No Known Activations