INDEX

Explanations

references to legal terminology or processes

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-0.41

 world

-0.34

 soeur

-0.33

-0.31

up

-0.31

 race

-0.31

sor

-0.30

-0.29

 capelli

-0.28

off

-0.28

POSITIVE LOGITS

 ویکی‌پدی

0.93

 kasarigan

0.85

 Italijanski

0.83

 autorytatywna

0.80

 Administrativna

0.79

 Wicidata

0.76

 beſte

0.76

ésultats

0.73

 increí

0.72

AndEndTag

0.72

Activations Density 0.000%

references to legal terminology or processes

No Comments

No Known Activations