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

 redes

-0.73

________________

-0.72

 conc

-0.67

Tex

-0.64

 Warn

-0.58

 cured

-0.57

 implanted

-0.57

 gems

-0.57

 impe

-0.56

 undefined

-0.56

POSITIVE LOGITS

noon

0.80

hedon

0.72

Tube

0.72

IPM

0.71

hran

0.70

inous

0.70

pires

0.69

horn

0.68

rupal

0.67

inating

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.