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

yip

-0.74

helle

-0.67

osate

-0.65

tracks

-0.64

aza

-0.64

iss

-0.63

ounter

-0.63

alogy

-0.63

scar

-0.63

 curs

-0.62

POSITIVE LOGITS

 Pompe

0.80

MENTS

0.72

 Salv

0.66

 Grade

0.66

 Owen

0.65

 Buildings

0.65

 Finch

0.63

 Cummings

0.62

 London

0.62

 Bean

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.