INDEX

Explanations

references to religious locations or activities

oai_token-act-pair · gpt-3.5-turbo

terms related to religious sites and practices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

estern

-0.76

 Carlson

-0.70

RAW

-0.68

balance

-0.67

urtle

-0.66

 Plex

-0.66

apse

-0.65

err

-0.63

NER

-0.61

olor

-0.61

POSITIVE LOGITS

 pilgrimage

1.28

 pilgrims

1.08

 maiden

1.01

 shrine

0.99

 pilgr

0.99

 worsh

0.88

 worshipped

0.83

 Shrine

0.83

 worship

0.82

antry

0.80

Activations Density 0.030%

references to religious locations or activities

terms related to religious sites and practices

No Comments

No Known Activations