INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ãĤ´ãĥ³

-0.92

baugh

-0.72

iquid

-0.70

ibu

-0.68

ãĥķ

-0.68

DeV

-0.68

MX

-0.66

trak

-0.66

Madison

-0.65

Eva

-0.65

POSITIVE LOGITS

idays

0.73

iversary

0.72

 Strongh

0.72

umbn

0.71

Ļ½

0.69

aucas

0.69

owship

0.68

ogle

0.66

ounty

0.65

 Recre

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.