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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

æĪ¦

-0.72

swick

-0.67

stood

-0.67

 Giants

-0.63

lain

-0.62

 Frie

-0.62

ãĤ´ãĥ³

-0.61

 sacrific

-0.61

 Britann

-0.60

 Radiant

-0.60

POSITIVE LOGITS

oday

0.66

livion

0.65

print

0.64

pace

0.63

 forg

0.63

LC

0.63

isine

0.62

igh

0.60

ichael

0.60

 âĸł

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.