INDEX

Explanations

references to physical exercises or activities

oai_token-act-pair · gpt-3.5-turbo

references to various types of "ups" or improvements

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

zi

-0.63

\\\\\\\\\\\\\\\\

-0.60

BN

-0.60

NB

-0.59

ty

-0.57

leasing

-0.56

OB

-0.56

FUL

-0.55

obi

-0.55

Bi

-0.54

POSITIVE LOGITS

hip

1.48

hips

1.43

ettings

1.29

paces

1.26

chool

1.26

uits

1.23

poons

1.21

pring

1.21

pace

1.21

ourcing

1.20

Activations Density 0.040%

references to physical exercises or activities

references to various types of "ups" or improvements

No Comments

No Known Activations