INDEX

Explanations

key identifiers and quality metrics in a dataset

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

numbers and names

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

 PLWABN

-0.60

@"/

-0.57

 فريبيس

-0.56

⤹

-0.53

󠁢

-0.50

internalType

-0.49

 CreateTagHelper

-0.49

 AssemblyCulture

-0.48

expandindo

-0.47

 Italijani

-0.47

POSITIVE LOGITS

VYMaps

0.44

 skolan

0.40

 Ejecutivo

0.39

ificados

0.38

izarse

0.38

 heureuse

0.37

guien

0.37

 تانيه

0.36

Джерела

0.36

 prisonniers

0.36

Activations Density 0.002%

key identifiers and quality metrics in a dataset

numbers and names

No Comments

No Known Activations