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

iatrics

-0.79

osate

-0.69

antine

-0.68

 Monk

-0.67

 referee

-0.67

 Scully

-0.65

 Eisen

-0.65

 Martial

-0.64

 Gund

-0.64

ths

-0.63

POSITIVE LOGITS

 period

0.96

 spending

0.94

period

0.91

 periods

0.90

 Period

0.75

inventoryQuantity

0.74

urally

0.72

xual

0.66

est

0.66

appropri

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.