INDEX

Explanations

mentions of various types of monks

oai_token-act-pair · gpt-3.5-turbo

mentions of monks and monasteries

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Houston

-0.79

 Orion

-0.78

 Hodg

-0.72

Allen

-0.71

Fle

-0.71

 Cyrus

-0.70

 Transportation

-0.70

Hash

-0.68

 Hash

-0.67

Spons

-0.66

POSITIVE LOGITS

 monk

3.32

 monks

3.24

 monastery

2.70

 nuns

2.49

 Monk

2.43

nun

2.31

 convent

1.86

 cleric

1.67

 clerics

1.63

 Abbey

1.57

Activations Density 0.032%

mentions of various types of monks

mentions of monks and monasteries

No Comments

No Known Activations