INDEX

Explanations

references to religious figures, specifically popes

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

Not in Any Lists

No Comments

Negative Logits

é¾

-0.84

ript

-0.83

yrinth

-0.76

hooting

-0.70

schild

-0.70

nel

-0.70

nesota

-0.70

Õ

-0.70

uld

-0.69

skirts

-0.69

POSITIVE LOGITS

 Francis

1.26

 Benedict

0.98

esses

0.90

 Clement

0.80

 Father

0.78

 infall

0.78

Father

0.78

otle

0.77

 Joseph

0.76

Franc

0.75

Activations Density 0.016%

references to religious figures, specifically popes

No Comments

No Known Activations