INDEX

Explanations

references to a specific character or entities named "Kerberos."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 33-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_33/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.33.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

 Fino

-0.41

 Dina

-0.39

造

-0.37

套

-0.37

 tụ

-0.36

intracht

-0.36

ANGO

-0.35

 moti

-0.35

ポスト

-0.35

tableFuture

-0.35

POSITIVE LOGITS

Ker

0.77

Ker

0.72

ker

0.69

KER

0.59

SourceChecksum

0.55

twimg

0.54

SuspendLayout

0.54

 فريبيس

0.53

ker

0.52

 препратки

0.48

Activations Density 0.179%

references to a specific character or entities named "Kerberos."

No Comments

No Known Activations