INDEX

Explanations

references to language and its various forms or contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

 ویکی‌پدی

-0.35

niosek

-0.33

 idéia

-0.32

➩

-0.32

Kết

-0.31

avax

-0.30

 ansehen

-0.29

 hindurch

-0.29

mejores

-0.28

 duquel

-0.28

POSITIVE LOGITS

auge

0.70

 arts

0.66

 Spoken

0.65

 Language

0.64

 Arts

0.64

 barrier

0.64

 spoken

0.63

0.62

anguage

0.62

 оригіналу

0.61

Activations Density 0.114%

references to language and its various forms or contexts

No Comments

No Known Activations