INDEX

Explanations

repeated references to "information."

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

ーキ

-0.42

 Leidenschaft

-0.40

 Behör

-0.40

 ilha

-0.39

 Fazit

-0.38

 dignité

-0.38

 estudar

-0.38

 infância

-0.38

 Befe

-0.37

 levure

-0.37

POSITIVE LOGITS

 information

2.03

Information

2.02

information

2.00

 Information

1.93

INFORMATION

1.69

 INFORMATION

1.68

informations

1.43

 información

1.39

 informations

1.32

 informatie

1.27

Activations Density 0.129%

repeated references to "information."

No Comments

No Known Activations