INDEX

Explanations

conditional phrases or situations

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

as in Notation

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

bootstrapcdn

-0.54

󠁢

-0.54

transQ

-0.54

MessageTagHelper

-0.52

SqlClient

-0.52

HasForeignKey

-0.50

posedge

-0.49

testify

-0.49

MLLoader

-0.48

MethodManager

-0.48

POSITIVE LOGITS

:✨

0.76

 nahilalakip

0.55

BeginInit

0.54

 propOrder

0.50

 nakalista

0.48

EndInit

0.41

CAPTION

0.39

Trama

0.39

 kasarigan

0.38

 autorytatywna

0.37

Activations Density 0.000%

conditional phrases or situations

as in Notation

No Comments

No Known Activations