INDEX

Explanations

references to various categories, particularly historical and architectural topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 15-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_15/width_16k/average_l0_78

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.15.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

RegressionTest

-0.71

ReusableCell

-0.69

AnchorStyles

-0.67

SourceChecksum

-0.67

CodeAttribute

-0.66

 protoimpl

-0.66

 surla

-0.65

writeField

-0.63

TagMode

-0.62

CloseOperation

-0.61

POSITIVE LOGITS

 north

0.58

 south

0.57

 jeunesse

0.57

 northeast

0.53

 southeast

0.53

olás

0.53

ool

0.52

 Norden

0.50

 female

0.50

northeast

0.50

Activations Density 0.563%

references to various categories, particularly historical and architectural topics

No Comments

No Known Activations