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

achus

-0.81

ĸļ

-0.80

udeb

-0.73

 Klux

-0.72

gradient

-0.70

malink

-0.70

hower

-0.70

andre

-0.70

escent

-0.69

sheet

-0.69

POSITIVE LOGITS

Proof

0.70

 pollen

0.67

DMV

0.65

Args

0.63

 arte

0.63

artifacts

0.62

 seizure

0.60

OF

0.60

 Nobel

0.60

HH

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.