INDEX

Explanations

mentions of religious and theological concepts, specifically focusing on questioning beliefs and actions related to God

oai_token-act-pair · gpt-3.5-turbo

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

TOR

-0.66

 Lima

-0.66

amina

-0.61

okemon

-0.61

"]=>

-0.61

ROR

-0.61

Wem

-0.59

 Indust

-0.59

 Sporting

-0.59

aukee

-0.59

POSITIVE LOGITS

frey

1.38

 Almighty

1.26

 Himself

1.24

father

1.21

 bless

1.14

mother

1.05

 Bless

1.04

win

1.04

 forbid

1.02

parents

0.97

Activations Density 0.025%

mentions of religious and theological concepts, specifically focusing on questioning beliefs and actions related to God

No Known Activations