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

tes

-0.74

pieces

-0.68

 restores

-0.67

marg

-0.64

 pills

-0.64

ful

-0.63

cond

-0.62

 inserts

-0.61

 Param

-0.60

tips

-0.60

POSITIVE LOGITS

KA

0.71

ILA

0.71

IDA

0.70

ilo

0.68

agle

0.67

ISE

0.65

UGH

0.65

ADRA

0.63

 Leilan

0.63

SIGN

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.