INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 notch

-0.72

 controls

-0.65

 nodd

-0.64

achev

-0.64

panel

-0.63

wick

-0.63

bridge

-0.62

 Wald

-0.62

ld

-0.61

aligned

-0.60

POSITIVE LOGITS

AME

0.72

 Rosa

0.68

 Sagan

0.66

 Sally

0.66

 Hobby

0.66

 Cosmos

0.66

ega

0.65

encia

0.65

yon

0.64

terday

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.