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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

elf

-0.74

 ILCS

-0.73

otom

-0.69

Sensor

-0.69

afety

-0.68

ibrary

-0.67

MpServer

-0.66

paren

-0.65

atoes

-0.64

Jr

-0.64

POSITIVE LOGITS

 Mant

0.78

 Conce

0.73

©¶æ¥µ

0.70

velt

0.67

 shrine

0.64

 Bland

0.64

Wem

0.63

 Darrell

0.63

 Lauren

0.61

 Nare

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.