INDEX

Explanations

phrases related to preventing or stopping someone or something from taking a specific action

oai_token-act-pair · gpt-3.5-turbo

references to preventing or stopping actions and behaviors

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

é¾įåĸļå£«

-0.88

=-=-

-0.68

rette

-0.64

entin

-0.64

soDeliveryDate

-0.64

estyles

-0.63

BAT

-0.62

ETHOD

-0.60

 Boone

-0.59

 Extras

-0.59

POSITIVE LOGITS

 from

1.08

from

0.94

 happening

0.92

 FROM

0.91

 harming

0.91

 slipping

0.85

 bleeding

0.82

 altogether

0.82

 accessing

0.81

 From

0.80

Activations Density 0.158%

phrases related to preventing or stopping someone or something from taking a specific action

references to preventing or stopping actions and behaviors

No Comments

No Known Activations