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

 Thrones

-0.78

fav

-0.73

 millenn

-0.71

Wi

-0.69

 adolesc

-0.69

aux

-0.69

 Dame

-0.67

dyn

-0.66

 intrigue

-0.65

 heel

-0.65

POSITIVE LOGITS

trump

0.82

orio

0.75

tracking

0.69

tracks

0.66

RI

0.66

owler

0.63

riber

0.63

nostic

0.62

£ı

0.62

aido

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.