INDEX

Explanations

terms related to victims or victimhood

oai_token-act-pair · gpt-3.5-turbo

references to victims of crimes or abuse

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

iannopoulos

-0.68

yip

-0.64

 Huck

-0.64

Ã¡s

-0.63

CLASSIFIED

-0.63

Owl

-0.61

adden

-0.60

leaf

-0.60

andel

-0.60

 snipp

-0.59

POSITIVE LOGITS

ization

0.93

izes

0.93

izers

0.91

istics

0.88

 blaming

0.88

izer

0.82

ize

0.79

hood

0.77

izations

0.77

izing

0.76

Activations Density 0.064%

terms related to victims or victimhood

references to victims of crimes or abuse

No Comments

No Known Activations