INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

mus

-0.76

mosp

-0.74

 Canad

-0.70

 revived

-0.68

united

-0.66

arag

-0.66

 Emirates

-0.64

aux

-0.64

 âĪ

-0.63

iversary

-0.63

POSITIVE LOGITS

hesda

1.13

terness

0.89

olulu

0.72

Zup

0.71

hower

0.69

 optimizing

0.65

 Hilbert

0.65

 cumbers

0.65

 Dise

0.62

andering

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.