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

iets

-0.83

lege

-0.78

spect

-0.73

itiveness

-0.72

FLAG

-0.68

æ©

-0.68

uces

-0.68

ignty

-0.68

ãĥĥãĥī

-0.67

minimum

-0.66

POSITIVE LOGITS

Neo

0.65

 christ

0.62

Oz

0.61

 disasters

0.59

 cout

0.59

 unknown

0.58

 being

0.58

 Eden

0.56

 abandonment

0.56

 Hell

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.