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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oÄŁ

-0.68

emin

-0.66

 Kislyak

-0.66

 Gork

-0.65

ornia

-0.63

gypt

-0.62

qqa

-0.62

 respons

-0.62

 Giuliani

-0.59

etsk

-0.58

POSITIVE LOGITS

ousy

0.77

 EntityItem

0.69

Osw

0.66

 Trick

0.66

iosyncr

0.64

ilk

0.64

OE

0.63

 Gleaming

0.62

abilia

0.62

ometown

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.