INDEX

Explanations

phrases expressing blame or disappointment towards someone or something

oai_token-act-pair · gpt-3.5-turbo

expressions of blame or shame

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

ccording

-0.78

endiary

-0.74

 teasp

-0.73

similar

-0.73

appropriately

-0.73

ultane

-0.69

amiya

-0.68

Sexual

-0.68

ONSORED

-0.68

roximately

-0.67

POSITIVE LOGITS

us

1.28

me

1.21

him

1.14

 thee

1.11

 them

1.08

ya

1.05

 yourselves

1.00

 THEM

0.92

 Them

0.92

Us

0.86

Activations Density 0.428%

phrases expressing blame or disappointment towards someone or something

expressions of blame or shame

No Comments

No Known Activations