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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

amara

-0.76

iscons

-0.72

ashtra

-0.68

Instance

-0.68

achable

-0.68

 plac

-0.67

earable

-0.65

ual

-0.65

ILLE

-0.62

natureconservancy

-0.61

POSITIVE LOGITS

matter

0.71

 antioxid

0.69

terday

0.68

 æľ

0.64

actionDate

0.63

\.

0.62

gay

0.62

msg

0.62

amy

0.61

 Outbreak

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.