INDEX

Explanations

phrases related to legal and political issues

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

 unemploy

-0.70

 looms

-0.64

-0.63

reon

-0.62

 spoil

-0.62

:/

-0.62

 combatants

-0.61

 entertained

-0.59

 nailed

-0.58

 lodged

-0.57

POSITIVE LOGITS

 addition

1.42

 spite

1.35

cluding

1.31

ventory

1.30

juries

1.30

herent

1.29

 contrast

1.28

jured

1.28

flation

1.24

 lieu

1.20

Activations Density 0.508%

phrases related to legal and political issues

No Comments

No Known Activations