INDEX

Explanations

references to confidentiality and information sharing

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

sharing secrets

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

igreja

-0.37

￣)

-0.36

 noqa

-0.36

)}</

-0.35

اما

-0.34

 EdgeInsets

-0.33

elfare

-0.33

 verlie

-0.33

 signal

-0.33

autant

-0.33

POSITIVE LOGITS

writeFieldEnd

0.70

SharedCtor

0.59

GenerationType

0.59

 secrets

0.58

 ModelExpression

0.58

secret

0.57

 secret

0.56

SECRET

0.55

 SECRET

0.54

CppMethod

0.54

Activations Density 0.113%

references to confidentiality and information sharing

sharing secrets

No Comments

No Known Activations