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

 brill

-0.81

sonian

-0.77

lov

-0.74

 communism

-0.72

cule

-0.69

 Communism

-0.68

aque

-0.68

 Chavez

-0.68

 toxin

-0.65

lys

-0.63

POSITIVE LOGITS

ows

1.25

owed

0.76

ãĥł

0.74

OWS

0.67

OW

0.67

xtap

0.67

orted

0.66

ategories

0.66

ower

0.66

owing

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.