INDEX

Explanations

ellipses or omissions in the text

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

show more

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

-------

-0.54

fromnode

-0.52

󠁢

-0.50

出版年

-0.47

AnchorStyles

-0.46

errHandler

-0.46

әрмәләр

-0.45

yarnpkg

-0.44

fillType

-0.44

nloa

-0.43

POSITIVE LOGITS

BufferException

0.57

 politiet

0.42

WebVitals

0.40

 trône

0.39

garek

0.39

 horaires

0.38

 témoignage

0.38

 cherchés

0.38

nezeu

0.38

 nalazi

0.37

Activations Density 0.030%

ellipses or omissions in the text

show more

No Comments

No Known Activations