INDEX

Explanations

the article "a" or "an" typically used before nouns

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

a [specific noun/adjective]

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

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

Viited

-0.31

 iirc

-0.26

 kebanyakan

-0.25

Notae

-0.25

stdarg

-0.25

 oddly

-0.23

 anderer

-0.23

жидан

-0.23

 plupart

-0.23

 pengar

-0.23

POSITIVE LOGITS

ChildScrollView

0.98

⟬

0.88

parsedMessage

0.75

 Administrativna

0.75

CreateModel

0.74

AddTagHelper

0.73

eleste

0.72

fillType

0.70

CreateMap

0.70

aibo

0.68

Activations Density 0.039%

the article "a" or "an" typically used before nouns

a [specific noun/adjective]

No Comments

No Known Activations