INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

icter

-0.77

geist

-0.73

 Extend

-0.69

 Â®

-0.67

 satir

-0.64

CVE

-0.64

 Mysteries

-0.63

 weap

-0.63

ãĤ±

-0.63

 fuse

-0.62

POSITIVE LOGITS

pter

0.86

cipled

0.82

ookie

0.80

cean

0.74

orthy

0.71

EStream

0.66

wt

0.65

WP

0.65

ocene

0.64

soDeliveryDate

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.