INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 ANGEL

-0.81

 Lear

-0.74

Gow

-0.71

GBT

-0.70

rane

-0.68

 Cree

-0.67

NESS

-0.66

 Fires

-0.64

 travelers

-0.63

oven

-0.62

POSITIVE LOGITS

anon

0.76

ographically

0.74

uably

0.70

successfully

0.69

dump

0.66

Stud

0.65

rehens

0.65

++++++++++++++++

0.65

ptin

0.64

verages

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.