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

uthor

-0.66

ilial

-0.66

tail

-0.64

 Caucasus

-0.64

oshenko

-0.63

 Tuls

-0.63

Bol

-0.62

Merit

-0.61

construct

-0.61

 Bulgar

-0.61

POSITIVE LOGITS

çīĪ

0.78

DH

0.68

itcher

0.65

ahime

0.65

 trough

0.65

 VIDEOS

0.63

MIA

0.61

emort

0.59

ilts

0.58

iary

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.