INDEX

Explanations

phrases related to legal actions or consequences, such as imposing fines, breaking antitrust rules, or needing a legal reason to do something

oai_token-act-pair · gpt-3.5-turbo

references to legal or regulatory actions against individuals or entities

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

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

 Letter

-0.64

DragonMagazine

-0.63

 Valentine

-0.60

 partName

-0.59

 Canaver

-0.59

enhagen

-0.58

 Newspaper

-0.57

 Dating

-0.57

netflix

-0.57

 Painter

-0.56

POSITIVE LOGITS

 theirs

0.86

 them

0.84

 goods

0.81

 ones

0.80

 products

0.76

 activity

0.75

%.

0.74

 wrongdoing

0.74

 thereto

0.72

 systems

0.72

Activations Density 0.955%

phrases related to legal actions or consequences, such as imposing fines, breaking antitrust rules, or needing a legal reason to do something

references to legal or regulatory actions against individuals or entities

No Comments

No Known Activations