INDEX

Explanations

sequences of characters comprising a specific non-English word with a particular pattern

oai_token-act-pair · gpt-3.5-turbo

non-standard or corrupted character sequences in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

onomy

-0.81

ortium

-0.76

erred

-0.67

ocamp

-0.67

osaurs

-0.66

 Grayson

-0.65

elope

-0.65

 subjected

-0.65

anwhile

-0.65

 polyg

-0.65

POSITIVE LOGITS

¹

1.60

º

1.56

³

1.55

²

1.52

¼

1.52

±

1.49

¶

1.37

´

1.36

¡

1.34

»

1.29

Activations Density 0.013%

sequences of characters comprising a specific non-English word with a particular pattern

non-standard or corrupted character sequences in text

No Comments

No Known Activations