INDEX

Explanations

mentions of people's names and titles

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

etheless

-0.87

atural

-0.84

ipeg

-0.83

hips

-0.80

iosity

-0.78

ionic

-0.75

ancial

-0.75

odd

-0.75

SpaceEngineers

-0.74

inatory

-0.74

POSITIVE LOGITS

lla

1.54

lli

1.51

Ã±o

1.32

cki

1.27

gger

1.26

llo

1.24

vich

1.22

lled

1.21

lda

1.19

ller

1.19

Activations Density 1.522%

mentions of people's names and titles

No Comments

No Known Activations