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

Catalog

-0.70

ļéĨĴ

-0.68

anche

-0.64

ĳå£«

-0.63

etsk

-0.61

 confinement

-0.59

 developer

-0.59

MpServer

-0.59

 assurances

-0.59

usra

-0.58

POSITIVE LOGITS

ensibly

0.78

ELF

0.78

 CARD

0.76

 mathemat

0.74

wagon

0.73

elf

0.69

haar

0.67

maxwell

0.67

ournal

0.65

agos

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.