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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

©¶æ

-0.72

NetMessage

-0.69

etheless

-0.67

 Rapt

-0.66

Fro

-0.65

odor

-0.64

ô

-0.64

 Elsa

-0.63

 Anita

-0.62

iflower

-0.62

POSITIVE LOGITS

 Scotch

0.74

MI

0.67

 depend

0.64

anie

0.64

 Highland

0.63

ACE

0.61

rocal

0.61

});

0.60

èĪ

0.60

 Associates

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.