INDEX

Explanations

phrases related to providing information or warnings to authorities or organizations

oai_token-act-pair · gpt-3.5-turbo

references to tips, knowledge, and information sharing related to awareness or forewarning

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

olars

-0.65

inals

-0.65

 Journals

-0.65

anco

-0.65

ï¸

-0.64

ocr

-0.64

aughs

-0.63

ogie

-0.63

inal

-0.62

eni

-0.62

POSITIVE LOGITS

 beforehand

0.78

worthiness

0.73

 alerts

0.67

 wrongdoing

0.67

ribune

0.66

ERROR

0.66

 purchaser

0.65

rosis

0.65

 distress

0.64

 suspicions

0.61

Activations Density 0.404%

phrases related to providing information or warnings to authorities or organizations

references to tips, knowledge, and information sharing related to awareness or forewarning

No Comments

No Known Activations