INDEX

Explanations

phrases related to someone being perceived as foolish or making mistakes

oai_token-act-pair · gpt-3.5-turbo

references to the concept of being a "fool."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Parenthood

-0.71

 Occupations

-0.68

lined

-0.66

 occupations

-0.60

accompan

-0.59

 contraction

-0.59

ccording

-0.59

 urgent

-0.58

 hotly

-0.57

Pwr

-0.56

POSITIVE LOGITS

proof

0.95

ishly

0.82

ery

0.80

hemer

0.79

ously

0.79

ibles

0.78

 fool

0.78

fusc

0.76

pas

0.76

ingly

0.76

Activations Density 0.013%

phrases related to someone being perceived as foolish or making mistakes

references to the concept of being a "fool."

No Comments

No Known Activations