INDEX

Explanations

phrases related to firearm companies, retirement, and political figures

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

 destro

-0.47

 challeng

-0.41

 Vaugh

-0.40

 disadvant

-0.36

 Azerb

-0.36

 conclud

-0.35

 agre

-0.33

 arrang

-0.33

 streng

-0.32

[*

-0.32

POSITIVE LOGITS

 âĢº

0.29

 Belfast

0.28

utterstock

0.27

ÂŃ

0.26

cigarettes

0.24

pin

0.23

Associated

0.23

heat

0.23

 Profile

0.23

 Copyright

0.23

Activations Density 11.657%

phrases related to firearm companies, retirement, and political figures

No Comments

No Known Activations