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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Abram

-0.73

abby

-0.68

 Luck

-0.68

 Upper

-0.67

imentary

-0.66

iflower

-0.65

 Dough

-0.63

 hemp

-0.62

pots

-0.62

 Lower

-0.62

POSITIVE LOGITS

ARC

0.70

tale

0.69

arde

0.68

isoft

0.65

EY

0.65

 corrid

0.64

 targ

0.64

gamer

0.62

 Clash

0.62

 exposition

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.