INDEX

Explanations

contractions of the words 'do not' or 'does not'

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "don’t" and its variations as indicators of negation or refusal

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 pione

-1.31

ò

-1.26

 eleph

-1.24

aditional

-1.23

Þ

-1.20

 practition

-1.19

ą

-1.18

ö

-1.18

é¾įå¥

-1.18

À

-1.18

POSITIVE LOGITS

't

1.61

Â´

0.89

\'

0.81

ned

0.80

ÃŃ

0.79

ï¿½

0.76

Ê

0.69

athan

0.67

uts

0.66

Activations Density 0.179%

contractions of the words 'do not' or 'does not'

instances of the phrase "don’t" and its variations as indicators of negation or refusal

No Comments

No Known Activations