INDEX

Explanations

phrases related to religious or theological concepts and themes

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Configuration

neuronpedia/gpt2-small__res_scl-ajt/6-res_scl-ajt

Prompts (Dashboard)

12,288 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

46,080

Data Type

torch.float32

Hook Point

blocks.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

apollo-research/Skylion007-openwebtext-tokenizer-gpt2

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hips

-0.83

 DARK

-0.72

cape

-0.72

 EntityItem

-0.71

folios

-0.71

 approve

-0.70

 Sacrament

-0.70

 Construct

-0.70

Constructed

-0.68

 Trader

-0.67

POSITIVE LOGITS

pty

1.42

mus

1.25

mers

1.23

orously

1.20

ulus

1.15

iday

1.15

bly

1.12

orous

1.10

mel

1.10

ming

1.09

Activations Density 7.892%

phrases related to religious or theological concepts and themes

No Comments

No Known Activations