INDEX

Explanations

numerical values and measurements related to dimensions and appearance

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

image filenames with extensions

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

glLoad

-0.34

LayoutStyle

-0.32

 StatefulWidget

-0.30

 fall

-0.30

 márm

-0.30

:+:

-0.30

 Maske

-0.30

javax

-0.29

 Stoffe

-0.29

principalTable

-0.28

POSITIVE LOGITS

 للمعارف

0.82

 propOrder

0.71

 autorytatywna

0.71

 ویکی‌پدی

0.65

addContainerGap

0.64

MessageOf

0.64

sizeCache

0.63

sidemargin

0.62

 invokingState

0.59

LookAnd

0.57

Activations Density 0.015%

numerical values and measurements related to dimensions and appearance

image filenames with extensions

No Comments

No Known Activations