INDEX

Explanations

phrases related to historical figures or religious saints

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

PT

-0.79

HO

-0.71

razil

-0.68

ORN

-0.64

itsch

-0.64

PN

-0.64

BIP

-0.63

clips

-0.62

OTH

-0.62

raltar

-0.61

POSITIVE LOGITS

 Laurent

1.05

 Clair

0.95

 Louis

0.95

 Augustine

0.94

 Lucia

0.93

 Petersburg

0.90

 Francis

0.88

clair

0.86

 Jude

0.81

onew

0.81

Activations Density 0.019%

phrases related to historical figures or religious saints

No Comments

No Known Activations