INDEX

Explanations

sequences of asterisks or dashes, indicating sections or separators in a document

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

separator lines

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

 aérea

-0.38

曖昧さ回避

-0.34

kháu

-0.34

 الرياضيه

-0.33

řeba

-0.33

wpi

-0.32

Datuak

-0.31

maybe

-0.29

 kuhusu

-0.28

 MwSt

-0.28

POSITIVE LOGITS

 CreateTagHelper

0.67

bibfield

0.59

BibitemShut

0.58

UnusedPrivate

0.56

SequentialGroup

0.55

 Paglinawan

0.55

 zwiſchen

0.54

 imagui

0.53

 Millennium

0.53

########.

0.53

Activations Density 0.013%

sequences of asterisks or dashes, indicating sections or separators in a document

separator lines

No Comments

No Known Activations