INDEX

Explanations

structured address or location formats, possibly related to legal or administrative content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/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.34.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

 downtown

-0.62

Downtown

-0.56

WriteBarrier

-0.55

downtown

-0.55

AndEndTag

-0.54

 externi

-0.52

 Downtown

-0.48

ьаж

-0.48

 gynhyrchwyd

-0.47

MarshalTo

-0.47

POSITIVE LOGITS

romyal

0.43

PopMatrix

0.42

EndContext

0.41

 PeEnEo

0.39

endipity

0.38

jamientos

0.36

Unternehmen

0.36

AutoModerator

0.35

﹍﹍﹍

0.35

especie

0.35

Activations Density 0.217%

structured address or location formats, possibly related to legal or administrative content

No Comments

No Known Activations