INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

seless

-0.67

omsday

-0.65

ubi

-0.64

Äĩ

-0.63

ebus

-0.63

wcs

-0.62

 Berger

-0.62

ww

-0.60

ORN

-0.60

"],"

-0.59

POSITIVE LOGITS

 Transactions

0.75

ethe

0.73

 Internal

0.72

agn

0.68

 Preferred

0.68

 disag

0.66

body

0.65

nesday

0.64

agnetic

0.62

naire

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.