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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

flat

-0.85

prototype

-0.81

pal

-0.79

pron

-0.76

orders

-0.74

Vari

-0.71

anim

-0.71

ador

-0.70

prints

-0.69

order

-0.66

POSITIVE LOGITS

OTAL

0.74

vernment

0.73

pheus

0.72

 slot

0.70

essee

0.70

utory

0.69

icum

0.65

 Torrent

0.64

bernatorial

0.63

atchewan

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.