INDEX

Explanations

numeric codes and identifiers related to data regions or structures

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

hexadecimal identifiers

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

 ujednoznacz

-0.37

InputTagHelper

-0.37

 fiches

-0.34

principalColumn

-0.33

 recipes

-0.33

nos

-0.31

}]{

-0.31

 recettes

-0.31

 inde

-0.30

 asequ

-0.30

POSITIVE LOGITS

setVerticalGroup

0.66

 betweenstory

0.65

astéro

0.64

 estekak

0.62

webElementXpaths

0.59

 Administrativna

0.56

SizeMode

0.54

featureID

0.53

Rohy

0.53

Дереккөздер

0.52

Activations Density 0.019%

numeric codes and identifiers related to data regions or structures

hexadecimal identifiers

No Comments

No Known Activations