INDEX

Explanations

terms related to comparisons or acknowledgments of positive qualities

oai_token-act-pair · gpt-3.5-turbo

expressions of respect and reverence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 spread

-0.67

 rack

-0.64

 spiral

-0.63

 trend

-0.63

obby

-0.61

 average

-0.58

 streak

-0.57

 slide

-0.56

Led

-0.55

spread

-0.55

POSITIVE LOGITS

 respects

4.03

 respect

1.98

 regards

1.91

 Respect

1.27

 respecting

1.27

respect

1.26

obe

1.20

 condolences

1.18

 respected

1.18

 compliments

1.17

Activations Density 0.007%

terms related to comparisons or acknowledgments of positive qualities

expressions of respect and reverence

No Comments

No Known Activations