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

 Logged

-0.79

 handling

-0.64

 Pearce

-0.63

 merch

-0.63

Posts

-0.62

 Handling

-0.61

 Gone

-0.60

eez

-0.58

 Estimated

-0.58

ocumented

-0.58

POSITIVE LOGITS

osphere

0.82

jri

0.78

heon

0.72

ricanes

0.70

æ©

0.68

cture

0.68

 Flavoring

0.68

ecake

0.66

okin

0.66

obin

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.