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

aukee

-0.69

 grizz

-0.67

anooga

-0.67

Roc

-0.67

 Grizz

-0.66

ither

-0.65

 Glac

-0.65

ationally

-0.64

 Pier

-0.63

 dentist

-0.63

POSITIVE LOGITS

åŃ

0.81

HTML

0.79

IPS

0.74

ROM

0.72

URI

0.71

WHO

0.70

CSS

0.70

CRIP

0.69

RFC

0.68

PDF

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.