INDEX

Explanations

elements in a structured data format, such as JSON or similar syntax

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

boolean and string types

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

AddTagHelper

-0.54

fromnode

-0.48

InitVars

-0.46

TargetException

-0.46



-0.46

ніципалі

-0.45

 AttributeSet

-0.45

richTextPanel

-0.45

AndEndTag

-0.44

LookAnd

-0.43

POSITIVE LOGITS

 bParam

0.48

yes

0.45

MethodType

0.42

BooleanField

0.40

 saites

0.39

yes

0.38

Prevention

0.38

是否

0.37

 становника

0.37

Yes

0.37

Activations Density 0.069%

elements in a structured data format, such as JSON or similar syntax

boolean and string types

No Comments

No Known Activations