INDEX

Explanations

terms and characters in Japanese and Korean contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_11/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

 تضيفلها

-0.69

setVerticalGroup

-0.63

httphttps

-0.59

uxxxx

-0.59

 ویکی‌پدی

-0.57

 <<<<<<<<<<<<<<

-0.55

 berke

-0.48

 snippetHide

-0.47

 setUsername

-0.47

 noDo

-0.47

POSITIVE LOGITS

Efq

0.57

 Houſe

0.46

CppMethod

0.45

 Majefty

0.43

 Anſ

0.41

 ivelany

0.41

 Cæsar

0.41

Зноскі

0.41

 houſe

0.40

 Pase

0.40

Activations Density 0.430%

terms and characters in Japanese and Korean contexts

No Comments

No Known Activations