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

 Canaver

-0.80

 looph

-0.73

 Flan

-0.72

McA

-0.68

htt

-0.68

 Bers

-0.66

 Schwar

-0.66

TDs

-0.64

 elim

-0.63

Kov

-0.63

POSITIVE LOGITS

å§«

0.83

bilt

0.78

UTION

0.76

Sov

0.71

aunder

0.70

ITNESS

0.69

Dust

0.69

Enjoy

0.68

ronic

0.68

edited

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.