INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

well

-0.71

anity

-0.70

tip

-0.69

 rundown

-0.66

archives

-0.65

 isolation

-0.64

andise

-0.63

cern

-0.62

 sidx

-0.62

oland

-0.62

POSITIVE LOGITS

ģĸ

0.75

HB

0.68

 Zoro

0.67

YP

0.65

enance

0.64

 Myst

0.64

 dele

0.64

ľ

0.63

 Debor

0.63

ropy

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.