INDEX

Explanations

phrases encouraging experimentation or trying new things

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

start then

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

ScopeManager

-0.53

 terciopelo

-0.43

 AssemblyCulture

-0.42

 contextLoads

-0.40

 escaleras

-0.39

viewDidLoad

-0.39

 aisladas

-0.39

 Cientí

-0.38

 annonces

-0.37

 informée

-0.37

POSITIVE LOGITS

 nahilalakip

0.64

 EconPapers

0.53

ůli

0.51

 invokingState

0.48

早速

0.46

 Norm

0.46

SequentialGroup

0.46

RTLR

0.46

AISSEE

0.45

invokeLater

0.45

Activations Density 0.024%

phrases encouraging experimentation or trying new things

start then

No Comments

No Known Activations