INDEX

Explanations

scientific measurements and numerical data, especially related to experimental procedures and results

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

doubling, halving, symmetry

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

LookAnd

-0.60

 فريبيس

-0.60

 autorytatywna

-0.59

setVerticalGroup

-0.59

 الرياضيه

-0.56

 مرئيه

-0.53

ReusableCell

-0.50

FunctionFlags

-0.50

LookAndFeel

-0.50

AddTagHelper

-0.49

POSITIVE LOGITS

 doubling

0.48

 doubled

0.45

一半

0.40

DOU

0.40

---|---|

0.40

 symmetrical

0.39

双

0.39

 twice

0.36

 Twice

0.36

 Symmetric

0.35

Activations Density 0.189%

scientific measurements and numerical data, especially related to experimental procedures and results

doubling, halving, symmetry

No Comments

No Known Activations