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

PRESS

-0.74

press

-0.73

âĸĪâĸĪ

-0.71

doi

-0.69

leck

-0.68

 Socrates

-0.68

ĸļ

-0.67

pressure

-0.66

 Greenpeace

-0.65

Wiki

-0.64

POSITIVE LOGITS

operator

0.68

anche

0.66

andowski

0.66

 execut

0.63

ydia

0.62

essee

0.61

ivid

0.60

 breaker

0.60

anced

0.60

cellaneous

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.