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

umbn

-0.71

oresc

-0.67

artney

-0.63

uncture

-0.63

airs

-0.63

icle

-0.61

Reloaded

-0.61

éĹĺ

-0.59

uder

-0.59

ighter

-0.58

POSITIVE LOGITS

 among

2.33

among

1.86

 amongst

1.81

 Among

1.25

Among

1.15

 between

0.66

 ILCS

0.64

»Ĵ

0.62

fman

0.61

 across

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.