INDEX

Explanations

numeric values or identifiers in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 13-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_13/width_65k/average_l0_75

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.13.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

 BorderSide

-0.56

Når

-0.54

 kasarigan

-0.54

StructEnd

-0.52

RTLR

-0.50

ResumeLayout

-0.50

alist

-0.49

èrement

-0.48

getField

-0.47

↵↵

-0.47

POSITIVE LOGITS

XmlAccessType

0.79

 فريبيس

0.70

 secondaires

0.69

HomeAsUpEnabled

0.69

 îna

0.63

0.62

 simplifié

0.60

VIAF

0.59

 mukana

0.57

 Ferner

0.57

Activations Density 0.002%

numeric values or identifiers in the text

No Comments

No Known Activations