INDEX

Explanations

legal documents and their formatting elements

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

final document pages

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

 للمعارف

-0.78

ſſung

-0.73

 مرئيه

-0.73

ロウィン

-0.71

 ſche

-0.71

 ſehr

-0.70

 ſei

-0.69

$_(

-0.69

ніципалі

-0.69

 zwiſchen

-0.68

POSITIVE LOGITS

indd

0.68

qxd

0.41

 last

0.32

 final

0.32

final

0.32

last

0.31

 letzter

0.27

revised

0.26

Last

0.26

ed

0.26

Activations Density 0.072%

legal documents and their formatting elements

final document pages

No Comments

No Known Activations