INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

timer

-0.72

coin

-0.71

 Lutheran

-0.70

lander

-0.66

 differe

-0.65

brew

-0.65

 shorth

-0.64

eas

-0.62

bench

-0.61

pass

-0.60

POSITIVE LOGITS

VERTIS

0.89

ailable

0.78

bernatorial

0.73

RANT

0.71

MIC

0.70

outheast

0.70

OME

0.70

 Dong

0.69

utra

0.69

illo

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.