INDEX

Explanations

phrases related to giving permission or requests for action

oai_token-act-pair · gpt-3.5-turbo

references to rules or guidelines about content sharing and attribution

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 srfAttach

-0.73

opers

-0.71

Ĥª

-0.70

 sensing

-0.68

Training

-0.67

rugged

-0.65

wagen

-0.65

grown

-0.63

Roaming

-0.63

STEM

-0.63

POSITIVE LOGITS

 publish

1.65

 reprint

1.59

 upload

1.32

 reprinted

1.28

 reproduce

1.21

 uploaded

1.21

 edit

1.18

 annot

1.16

 archive

1.16

 publishing

1.14

Activations Density 0.460%

phrases related to giving permission or requests for action

references to rules or guidelines about content sharing and attribution

No Comments

No Known Activations