INDEX

Explanations

alert or warning messages related to safety or functionality

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.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

 únicos

-0.41

 camada

-0.40

appartamento

-0.39

Londres

-0.39

 montanha

-0.39

 exercí

-0.39

 praktik

-0.38

sbereich

-0.37

říklad

-0.36

 Unternehmens

-0.36

POSITIVE LOGITS

 alert

1.35

 Alert

1.27

alert

1.27

 alarm

1.23

Alert

1.21

meter

1.17

 meter

1.13

 Meter

1.11

alarm

1.09

Meter

1.09

Activations Density 0.116%

alert or warning messages related to safety or functionality

No Comments

No Known Activations