INDEX

Explanations

specific numerical values and expected ranges related to measurements or predictions

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

numbers and measurements

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

整

-0.34

 coin

-0.33

放

-0.33

HIR

-0.31

]]]

-0.31

 posting

-0.31

 }}">

-0.30

➊

-0.30

teras

-0.30

整

-0.29

POSITIVE LOGITS

IsMutable

0.62

AsUp

0.62

HasBeenSet

0.52

ніципа

0.52

AntiForgeryToken

0.51

 hObject

0.49

NewUrlParser

0.48

 فريبيس

0.48

 recommandée

0.48

 propOrder

0.47

Activations Density 0.097%

specific numerical values and expected ranges related to measurements or predictions

numbers and measurements

No Comments

No Known Activations