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

aband

-0.79

adv

-0.68

à©

-0.59

 ours

-0.58

 swoop

-0.58

 battleground

-0.58

 rails

-0.58

NetMessage

-0.57

commerce

-0.57

built

-0.57

POSITIVE LOGITS

ttle

0.82

icter

0.75

hiba

0.74

fit

0.71

ogen

0.71

 outwe

0.70

atron

0.68

ople

0.67

ELY

0.67

enium

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.