INDEX

Explanations

structured data related to scientific or academic content

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

 autorytatywna

-0.52

unduh

-0.44

AccessorTable

-0.44

upsi

-0.43

 camiset

-0.43

hilangan

-0.42

 resourceCulture

-0.41

RegressionTest

-0.40

 endblock

-0.39

aperto

-0.38

POSITIVE LOGITS

TestingModule

0.48

 kaarangay

0.42

 ffilmiau

0.39

تقاوى

0.39

 säkert

0.38

vinfos

0.37

represent

0.36

 diagrams

0.36

 represent

0.36

tgär

0.36

Activations Density 0.301%

structured data related to scientific or academic content

No Comments

No Known Activations