INDEX

Explanations

phrases related to political news and subscription services

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

 Unlimited

-0.62

 Mortal

-0.58

 doom

-0.56

uthor

-0.54

 Nero

-0.54

 dispute

-0.54

iba

-0.53

Nun

-0.53

 Kills

-0.53

 clone

-0.53

POSITIVE LOGITS

Swe

0.76

åĤ

0.72

å·

0.70

chairs

0.70

psc

0.70

à¼

0.69

éļ

0.68

atalie

0.68

Afee

0.67

Â¯

0.67

Activations Density 0.021%

phrases related to political news and subscription services

No Comments

No Known Activations