INDEX

Explanations

words related to negative or harmful actions or behaviors

oai_token-act-pair · gpt-3.5-turbo

references to negative behaviors and derogatory remarks

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

alg

-0.70

atorium

-0.64

amaz

-0.63

cture

-0.63

apest

-0.62

ouple

-0.62

natureconservancy

-0.62

idated

-0.61

00200000

-0.61

 horm

-0.59

POSITIVE LOGITS

 against

1.07

 unlawfully

0.98

 towards

0.97

 whilst

0.95

 toward

0.94

 throughout

0.92

 during

0.92

 indiscrim

0.91

 inappropriately

0.89

 illegally

0.89

Activations Density 0.447%

words related to negative or harmful actions or behaviors

references to negative behaviors and derogatory remarks

No Comments

No Known Activations