INDEX

Explanations

references to numerical data or formatting in code

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

parameter or code formatting

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

tagHelper

-0.54

archiviato

-0.43

makeConstraints

-0.43

Hozzáférés

-0.42

IsContent

-0.41

 wikipagina

-0.41

TestingModule

-0.41

awaiter

-0.40

DIPSETTING

-0.38

iprot

-0.37

POSITIVE LOGITS

complexContent

0.56

 isNameExpr

0.56

Seni

0.51

quatch

0.51

Nino

0.50

ArgsConstructor

0.49

Nimbus

0.48

SJ

0.48

 EconPapers

0.48

hasMoreElements

0.47

Activations Density 0.005%

references to numerical data or formatting in code

parameter or code formatting

No Comments

No Known Activations