INDEX

Explanations

text related to providing guidance or explanations

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate guidance or instructional content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Virgin

-0.72

ussia

-0.70

 Arabs

-0.69

ãĥ©ãĥ³

-0.68

Jews

-0.67

Islamic

-0.66

Muslims

-0.65

Nazis

-0.64

ansas

-0.63

 disapprove

-0.62

POSITIVE LOGITS

 infographic

0.90

Fortunately

0.87

Luckily

0.83

 Luckily

0.83

ado

0.78

 Fortunately

0.77

Thankfully

0.76

 luckily

0.76

 Thankfully

0.75

 fortunately

0.74

Activations Density 0.555%

text related to providing guidance or explanations

phrases that indicate guidance or instructional content

No Comments

No Known Activations