INDEX

Explanations

concepts related to sorting and categorization

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

indicate whether

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

SequentialGroup

-0.51

Gail

-0.50

 tartalomajánló

-0.49

ranton

-0.49

bkz

-0.48

lao

-0.48

laught

-0.47

"?>

-0.47

dene

-0.47

gai

-0.46

POSITIVE LOGITS

oredCriteria

0.49

DockStyle

0.49

 pinulongan

0.47

исленность

0.44

LabelTagHelper

0.40

 OMITBAD

0.39

 Infórmanos

0.39

AndEndTag

0.38

 froid

0.37

 muestre

0.36

Activations Density 0.008%

concepts related to sorting and categorization

indicate whether

No Comments

No Known Activations

concepts related to sorting and categorization

indicate whether

No Comments

No Known Activations