INDEX

Explanations

mentions of things being open or available for access or participation

oai_token-act-pair · gpt-3.5-turbo

instances of the word "open" and its variations within different contexts

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

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

ILY

-0.67

 Deaths

-0.67

 derives

-0.66

MENTS

-0.65

anium

-0.64

 Surv

-0.63

eree

-0.63

MENT

-0.63

yon

-0.60

parents

-0.60

POSITIVE LOGITS

 open

3.59

open

2.66

Open

2.22

 Open

2.21

 OPEN

2.18

 opens

1.83

 opened

1.80

 openness

1.75

 closed

1.71

 reopen

1.64

Activations Density 0.022%

mentions of things being open or available for access or participation

instances of the word "open" and its variations within different contexts

No Comments

No Known Activations