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

alty

-0.69

ALS

-0.69

York

-0.68

otos

-0.67

aux

-0.66

ongh

-0.65

axe

-0.64

Hamilton

-0.64

 Bethlehem

-0.63

 metic

-0.62

POSITIVE LOGITS

 period

0.99

 Party

0.87

 periods

0.77

ioxide

0.75

strate

0.69

utterstock

0.68

 Parenthood

0.66

period

0.65

esi

0.64

 Parties

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.