INDEX

Explanations

words related to technology and programming

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Rasmussen

-0.90

udence

-0.80

 undercover

-0.80

jud

-0.77

conduct

-0.75

 Reeves

-0.73

roit

-0.72

lethal

-0.72

stre

-0.71

POSITIVE LOGITS

 Linux

1.47

 macOS

1.34

 Unix

1.32

 Windows

1.31

 interoper

1.30

Linux

1.30

 desktop

1.29

GNU

1.27

 linux

1.27

 Desktop

1.25

Activations Density 0.600%

words related to technology and programming

No Comments

No Known Activations