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

velt

-0.80

letes

-0.73

 Drac

-0.72

aug

-0.71

juven

-0.71

apego

-0.68

arts

-0.66

helm

-0.66

imar

-0.65

 Frey

-0.65

POSITIVE LOGITS

ING

0.67

 dreaded

0.66

ASH

0.64

ª

0.63

 disadvantage

0.61

 landfill

0.61

MI

0.61

 coveted

0.60

 dearly

0.60

 lucrative

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.