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

astery

-0.83

pine

-0.81

atin

-0.74

angan

-0.72

ģĸ

-0.72

 lockout

-0.67

 kettle

-0.66

 Practices

-0.66

 blocker

-0.66

anke

-0.65

POSITIVE LOGITS

stab

0.71

åĤ

0.71

Ò

0.69

 flowers

0.66

+.

0.66

birds

0.66

;;;;

0.66

#$#$

0.64

 Egyptians

0.61

ublic

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.