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

*/(

-0.88

 Flavoring

-0.85

tails

-0.77

 guiActiveUnfocused

-0.71

 Editorial

-0.65

ASON

-0.65

PLA

-0.65

boards

-0.64

 ãĤµ

-0.63

Boss

-0.63

POSITIVE LOGITS

 child

1.03

age

1.03

 ages

1.02

child

0.81

 Child

0.74

>>\

0.72

abase

0.71

rape

0.70

 Carnegie

0.65

 euth

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.