INDEX

Explanations

passive voice constructions with "being" followed by a verb

oai_token-act-pair · gpt-3.5-turbo

instances of the word "being" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ismo

-0.73

 Explosion

-0.73

 constellation

-0.67

 Prescott

-0.64

 heavens

-0.63

lap

-0.60

inav

-0.60

 Farming

-0.59

itsch

-0.59

 Philippines

-0.59

POSITIVE LOGITS

 able

1.18

 judged

0.87

 replaced

0.83

 eaten

0.81

 aware

0.80

 subjected

0.78

Ĥ¬

0.78

ĵĺ

0.77

 considered

0.77

held

0.76

Activations Density 0.064%

passive voice constructions with "being" followed by a verb

instances of the word "being" in various contexts

No Comments

No Known Activations