INDEX

Explanations

detailed information such as numbers with decimal points

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 piping

-0.78

 decor

-0.71

 unlucky

-0.68

 sabotage

-0.66

 downt

-0.65

manship

-0.65

 loyalty

-0.64

 revenge

-0.63

 symm

-0.63

 victories

-0.63

POSITIVE LOGITS

doi

1.50

doi

1.40

DOI

1.29

 PubMed

1.25

 Accessed

1.22

 ISBN

1.21

 Retrieved

1.13

Abstract

1.12

âĵĺ

1.10

 Available

1.07

Activations Density 0.082%

detailed information such as numbers with decimal points

No Comments

No Known Activations