INDEX

Explanations

punctuation marks and their contexts in the text

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

end of sentence punctuation

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

ngdoc

-0.65

contentLoaded

-0.59

 AssemblyCompany

-0.59

AddTagHelper

-0.58

 Мексичка

-0.57

 متحده

-0.57

 estekak

-0.57

AnchorTagHelper

-0.57

AndEndTag

-0.56

 bakgrund

-0.56

POSITIVE LOGITS

or

0.37

 especially

0.31

änden

0.30

IF

0.30

Bis

0.28

 eller

0.28

HP

0.27

RVA

0.27

 Especially

0.27

especially

0.27

Activations Density 0.143%

punctuation marks and their contexts in the text

end of sentence punctuation

No Comments

No Known Activations