INDEX

Explanations

proper names, specifically names of individuals

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

 Uncharted

-0.76

 Wilmington

-0.74

 Californ

-0.70

 Colossus

-0.69

 Phillies

-0.68

 Bottle

-0.66

 Wilde

-0.65

 ster

-0.65

 Gravity

-0.65

ppel

-0.65

POSITIVE LOGITS

 Hasan

1.37

uddin

1.37

 Hassan

1.29

 Ahmad

1.27

 Muhammad

1.23

Ali

1.22

iyah

1.20

 Khan

1.20

 Mahm

1.20

 Sheikh

1.17

Activations Density 1.081%

proper names, specifically names of individuals

No Comments

No Known Activations