INDEX

Explanations

names related to individuals in different contexts

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

anwhile

-0.35

 Vaugh

-0.31

 destro

-0.31

'."

-0.30

 conclud

-0.30

 disadvant

-0.30

enegger

-0.29

 Azerb

-0.28

 furthermore

-0.27

 helicop

-0.26

POSITIVE LOGITS

 âĢº

0.35

taboola

0.28

):

0.27

)--

0.27

\":

0.27

 ..........

0.26

utterstock

0.25

natureconservancy

0.24

 Belfast

0.24

 âĢ¢

0.23

Activations Density 15.370%

names related to individuals in different contexts

No Comments

No Known Activations