INDEX

Explanations

actions or behaviors that are morally or legally questionable

oai_token-act-pair · gpt-3.5-turbo

references to compliance and misconduct in professional conduct

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

Luck

-0.77

 aging

-0.75

Tuc

-0.74

 Jelly

-0.73

 ageing

-0.71

 rebuilt

-0.69

 Dying

-0.68

 Pigs

-0.67

 reborn

-0.67

 Aging

-0.66

POSITIVE LOGITS

 violate

1.38

 endanger

1.32

 jeopard

1.31

 impair

1.23

 adversely

1.18

 incrim

1.13

 contra

1.11

rouse

1.10

 violates

1.10

 prejud

1.09

Activations Density 0.375%

actions or behaviors that are morally or legally questionable

references to compliance and misconduct in professional conduct

No Comments

No Known Activations