INDEX

Explanations

instances of the word "lying"

oai_token-act-pair · gpt-3.5-turbo

instances and discussions of dishonesty or deception

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Gust

-0.73

 Generations

-0.67

 Flavor

-0.67

Pacific

-0.67

anan

-0.66

cend

-0.65

ender

-0.62

ETF

-0.61

ADE

-0.61

 Pione

-0.61

POSITIVE LOGITS

 lying

3.78

lie

1.99

 lies

1.80

 laying

1.62

 liar

1.60

 lied

1.57

 sitting

1.41

 kneeling

1.38

 hiding

1.37

 Lies

1.26

Activations Density 0.014%

instances of the word "lying"

instances and discussions of dishonesty or deception

No Comments

No Known Activations