INDEX

Explanations

numbers and symbols such as characters for specific syntax and formatting

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

 ropes

-0.76

 Agric

-0.73

 seiz

-0.73

Gan

-0.68

 horizont

-0.67

 Palest

-0.65

çīĪ

-0.63

 Marshal

-0.63

 Mechdragon

-0.63

 Indies

-0.62

POSITIVE LOGITS

Ĵ

1.08

º

1.03

ķ

1.02

Â¢

1.00

¹

0.99

¡

0.98

ª

0.97

£

0.94

®

0.93

¬

0.92

Activations Density 0.091%

numbers and symbols such as characters for specific syntax and formatting

No Comments

No Known Activations