INDEX

Explanations

website links

oai_token-act-pair · gpt-3.5-turbo

punctuation marks and sentence structures in the document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 commodities

-0.80

 attrition

-0.76

 extingu

-0.71

 extinguished

-0.70

 consum

-0.69

 cumulative

-0.68

 expended

-0.67

 gases

-0.67

 curfew

-0.67

 diminishing

-0.66

POSITIVE LOGITS

 Basically

1.39

 Apparently

1.34

 Seems

1.27

Its

1.27

It

1.25

 Anyway

1.21

 Seriously

1.21

 Essentially

1.16

 Honestly

1.15

 Turns

1.15

Activations Density 0.379%

website links

punctuation marks and sentence structures in the document

No Comments

No Known Activations