INDEX

Explanations

references to established principles or beliefs, particularly focusing on doctrines

oai_token-act-pair · gpt-3.5-turbo

references to doctrinal concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ells

-0.73

vals

-0.67

 Cheong

-0.66

RED

-0.65

BBC

-0.64

 Democr

-0.64

RGB

-0.63

PU

-0.63

ALTH

-0.63

POR

-0.62

POSITIVE LOGITS

 doctrine

1.46

 Doctrine

1.26

 doctrines

1.07

istries

0.86

 theology

0.80

 guiActiveUn

0.79

 doctr

0.78

ieu

0.75

illy

0.73

ervative

0.72

Activations Density 0.008%

references to established principles or beliefs, particularly focusing on doctrines

references to doctrinal concepts

No Comments

No Known Activations