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

ridges

-0.74

mpeg

-0.73

 Quan

-0.72

 Dunk

-0.70

igham

-0.68

nings

-0.67

emaker

-0.66

 Rafael

-0.65

rike

-0.65

 Noct

-0.65

POSITIVE LOGITS

éĹ

0.69

channelAvailability

0.66

Uncommon

0.63

CHAR

0.61

CENT

0.61

archy

0.59

=~

0.59

 abound

0.59

FREE

0.58

stars

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.