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

zai

-0.68

terness

-0.67

 looting

-0.67

icipated

-0.66

comings

-0.65

 presentations

-0.65

 cooperating

-0.63

metics

-0.62

 Angela

-0.62

jong

-0.60

POSITIVE LOGITS

alt

1.90

ALT

0.98

ain

0.84

ieth

0.71

hide

0.69

ãĤĭ

0.68

ickets

0.67

earth

0.65

isec

0.65

anson

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.