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

 Gunn

-0.69

otine

-0.68

 Watson

-0.66

 Scouts

-0.65

ovember

-0.64

igham

-0.63

 Patri

-0.63

 Sharma

-0.63

 Regina

-0.63

wx

-0.62

POSITIVE LOGITS

utations

0.66

deal

0.65

maximum

0.65

veyard

0.65

account

0.64

 ç¥ŀ

0.63

closed

0.61

©¶æ

0.60

 Locked

0.60

ruction

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.