INDEX

Explanations

instances of specific characters or people in a narrative context

oai_token-act-pair · gpt-4o-mini Triggered by @bot

years ago

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

The neuron detects references to people — named persons, titles, and personal pronouns (mentions of individuals like "defendant," "he," "she," names, and honorifics).

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

AsUp

-0.50

 समीक्षक

-0.49

 tabPage

-0.45

 voet

-0.42

 <<<<<<<<<<<<<<

-0.40

FindBy

-0.39

 znamen

-0.38

PDATE

-0.38

Subview

-0.38

 gebo

-0.38

POSITIVE LOGITS

:✨

0.47

GenerationType

0.39

ragalactic

0.39

TagMode

0.38

romechanical

0.38

 repeatedly

0.37

 covenants

0.37

MockBean

0.36

 beginnetje

0.35

 صوتيه

0.35

Activations Density 0.083%

instances of specific characters or people in a narrative context

years ago

The neuron detects references to people — named persons, titles, and personal pronouns (mentions of individuals like "defendant," "he," "she," names, and honorifics).

No Comments

No Known Activations

instances of specific characters or people in a narrative context

years ago

The neuron detects references to people — named persons, titles, and personal pronouns (mentions of individuals like "defendant," "he," "she," names, and honorifics).

No Comments

No Known Activations