INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 infield

-0.72

endif

-0.64

adan

-0.60

 Idol

-0.60

 adjective

-0.60

aux

-0.59

 toes

-0.59

wcsstore

-0.58

*/(

-0.58

 microphone

-0.58

POSITIVE LOGITS

lees

0.72

ONT

0.71

RECT

0.69

ONSORED

0.67

 Fract

0.66

encing

0.66

rers

0.65

ements

0.65

ENTS

0.65

onew

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.