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

 tremend

-0.81

 hemor

-0.80

 Savings

-0.75

 Annotations

-0.71

Denver

-0.68

 Newsp

-0.67

 fare

-0.64

 Junk

-0.64

Benef

-0.62

 Shack

-0.62

POSITIVE LOGITS

ion

1.68

area

0.82

otiation

0.78

iage

0.73

age

0.72

È

0.71

ioned

0.68

aged

0.67

Ã³

0.67

tesy

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.