INDEX

Explanations

phrases related to political actions and international relations, with a special focus on government decisions and policies

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 scenery

-0.74

izont

-0.71

 everywhere

-0.69

itably

-0.69

ishly

-0.68

 sleeve

-0.67

 fishes

-0.65

isable

-0.64

à¨

-0.64

 ourselves

-0.64

POSITIVE LOGITS

 However

1.26

 Critics

1.25

 Earlier

1.22

 Previously

1.19

 Additionally

1.14

 Currently

1.13

 Since

1.13

 Though

1.12

 Although

1.12

 Initially

1.09

Activations Density 0.437%

phrases related to political actions and international relations, with a special focus on government decisions and policies

No Comments

No Known Activations