INDEX

Explanations

occurrences of the word "end."

oai_token-act-pair · gpt-3.5-turbo

references to endings or conclusions in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Cosponsors

-0.87

é¾įåĸļå£«

-0.75

uyomi

-0.71

»Ĵ

-0.69

Role

-0.69

externalActionCode

-0.68

leground

-0.67

 è£ıè¦ļéĨĴ

-0.67

 guiIcon

-0.66

NC

-0.65

POSITIVE LOGITS

 humankind

0.76

 eternity

0.76

 mankind

0.75

onement

0.73

 reckoning

0.71

 humanity

0.70

 course

0.69

 boredom

0.68

 separation

0.68

 hostilities

0.68

Activations Density 0.115%

occurrences of the word "end."

references to endings or conclusions in various contexts

No Comments

No Known Activations