INDEX

Explanations

quantitative comparisons regarding amounts

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

 sumpay

-0.68

 ddelwed

-0.66

 Paglinawan

-0.66

ſſung

-0.65

ésultats

-0.64

 vooz

-0.64

 ſans

-0.63

iſche

-0.62

cedures

-0.62

enablog

-0.62

POSITIVE LOGITS

0.42

in

0.40

 again

0.37

 what

0.36

 about

0.36

 because

0.35

 else

0.35

again

0.34

for

0.33

mapStateToProps

0.33

Activations Density 0.021%

quantitative comparisons regarding amounts

No Comments

No Known Activations