INDEX

Explanations

instances of illegal or unethical activities or behaviors

oai_token-act-pair · gpt-3.5-turbo

instances of harassment and assault

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

é¾įåĸļå£«

-0.94

inventoryQuantity

-0.90

ãĤ´ãĥ³

-0.84

terms

-0.82

wcsstore

-0.81

uer

-0.78

itsch

-0.78

urgy

-0.78

mpeg

-0.77

 Subtle

-0.77

POSITIVE LOGITS

 passers

1.47

 unsuspecting

1.31

 unwitting

1.30

 volunteers

1.22

 residents

1.19

 individuals

1.17

 participants

1.17

 attendees

1.16

 informants

1.16

 activists

1.15

Activations Density 0.402%

instances of illegal or unethical activities or behaviors

instances of harassment and assault

No Comments

No Known Activations