INDEX

Explanations

tables or structured data representations

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

table separators or headers

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

dAtA

-0.54

 thereafter

-0.45

CDCl

-0.43

 toJson

-0.42

 dalamnya

-0.42

principalColumn

-0.42

tinyos

-0.40

 hopefully

-0.40

AsUp

-0.40

 luckily

-0.40

POSITIVE LOGITS

 تضيفلها

0.61

colhead

0.59

__(/*!

0.53

 cherchés

0.53

<th>

0.53

脚注の使い方

0.51

 name

0.49

 তারিখ

0.49

 voks

0.46

Name

0.44

Activations Density 0.011%

tables or structured data representations

table separators or headers

No Comments

No Known Activations