INDEX

Explanations

instances of numerical or categorical cases within a specified context

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

case followed by string literals

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

 mourut

-0.39

 ciudadana

-0.37

 sentiers

-0.35

 muerto

-0.35

@[+][

-0.34

exels

-0.33

 světě

-0.33

readValue

-0.32

 sendero

-0.32

 černá

-0.32

POSITIVE LOGITS

Tembelea

0.64

ChildScrollView

0.63

期刊论文

0.54

CppMethod

0.54

 estekak

0.54

⎛

0.53

usercontent

0.53

roni

0.53

MigrationBuilder

0.52

<sup>

0.52

Activations Density 0.001%

instances of numerical or categorical cases within a specified context

case followed by string literals

No Comments

No Known Activations