INDEX

Explanations

political terms and phrases related to the Left

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

andise

-0.80

alez

-0.79

è¦ļéĨĴ

-0.73

ãĤ¼ãĤ¦ãĤ¹

-0.71

 TOUR

-0.70

ãĥīãĥ©

-0.68

 Allaah

-0.67

CLASSIFIED

-0.65

 Mehran

-0.63

Delivery

-0.63

POSITIVE LOGITS

wing

1.32

overs

1.27

 wing

1.10

 flank

1.04

ward

1.02

leaning

0.93

 sided

0.91

 leaning

0.90

wich

0.90

most

0.83

Activations Density 0.039%

political terms and phrases related to the Left

No Comments

No Known Activations