INDEX

Explanations

phrases that suggest large quantities, positive evaluations, and specific characteristics or comparisons

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

degree qualifiers

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

Personendaten

-0.48

and

-0.46

 selain

-0.36

 Töchter

-0.34

 GoogleFonts

-0.34

debugging

-0.34

-0.31

 FontFamily

-0.31

cookieParser

-0.30

HasForeignKey

-0.30

POSITIVE LOGITS

+#+#

0.71

 protoimpl

0.56

 تضيفلها

0.54

évaluateur

0.54

esgue

0.52

httphttps

0.51

Tembelea

0.51

########.

0.50

usercontent

0.50

Portail

0.48

Activations Density 0.055%

phrases that suggest large quantities, positive evaluations, and specific characteristics or comparisons

degree qualifiers

No Comments

No Known Activations