INDEX

Explanations

phrases indicating the introduction or specification of content

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

the following

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

 الرياضيه

-0.46

 Italijanski

-0.44

 bezeichneter

-0.44

+:+

-0.44

]=>

-0.43

 hikariConfig

-0.40

 Мексичка

-0.40

原始内容存档于

-0.38

Cyfarwyddwr

-0.37

CROSSTALK

-0.36

POSITIVE LOGITS

lüğ

0.56

 ſeines

0.56

 surla

0.55

出版年

0.55

 dieſen

0.54

cés

0.53

 queſto

0.53

<unused10>

0.53

 好文分享

0.53

<unused30>

0.53

Activations Density 0.027%

phrases indicating the introduction or specification of content

the following

No Comments

No Known Activations