INDEX

Explanations

phrases related to political discourse and policy 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

xtap

-0.91

atan

-0.76

DragonMagazine

-0.69

IPS

-0.66

OCK

-0.65

ĺħ

-0.64

uca

-0.64

uckle

-0.63

ovie

-0.63

Enlarge

-0.61

POSITIVE LOGITS

 than

1.55

 importantly

1.20

 Than

1.05

than

1.00

 stringent

0.97

ado

0.93

 sophisticated

0.91

 expensive

0.91

 complicated

0.87

 efficient

0.87

Activations Density 0.371%

phrases related to political discourse and policy issues

No Comments

No Known Activations