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

essee

-0.70

illery

-0.69

 Gators

-0.69

uras

-0.68

 illiter

-0.66

ucky

-0.66

 Kentucky

-0.65

enza

-0.65

basketball

-0.64

Michigan

-0.64

POSITIVE LOGITS

 trailing

0.77

actionDate

0.76

¿½

0.75

cknow

0.72

Ñĭ

0.68

 partName

0.66

ãĤ©

0.65

 Medline

0.65

âķĲ

0.65

 pregn

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.