INDEX

Explanations

words related to negative impact or disruption

oai_token-act-pair · gpt-3.5-turbo

words and phrases associated with causing harm or negative consequences

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

"},

-0.65

 Seeking

-0.60

 spons

-0.59

Activity

-0.58

bleacher

-0.57

nings

-0.57

 regards

-0.56

ordering

-0.56

DragonMagazine

-0.55

 rooting

-0.55

POSITIVE LOGITS

ively

1.15

 yourselves

1.03

 herself

1.00

 himself

0.99

 yourself

0.88

iously

0.88

 themselves

0.88

ingly

0.85

ously

0.83

 ourselves

0.83

Activations Density 0.486%

words related to negative impact or disruption

words and phrases associated with causing harm or negative consequences

No Comments

No Known Activations