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

Leaks

-0.93

idon

-0.67

ricting

-0.65

LCS

-0.65

doms

-0.64

GD

-0.63

LS

-0.63

TPS

-0.62

Dru

-0.62

 GMOs

-0.60

POSITIVE LOGITS

BuyableInstoreAndOnline

1.08

ansky

0.78

intage

0.75

english

0.71

usa

0.71

 earthly

0.70

itime

0.68

ummer

0.67

luster

0.67

 Wichita

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.