INDEX

Explanations

phone calls and medical profession-related terms

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

ĪĴ

-0.89

ãĥĹ

-0.65

 Yamaha

-0.62

 contagious

-0.62

Ĥ¬

-0.61

 juggling

-0.61

 Maker

-0.58

 dish

-0.56

 mornings

-0.56

Magikarp

-0.56

POSITIVE LOGITS

ologne

1.31

redits

1.26

ategories

1.24

ounters

1.21

oding

1.20

ursor

1.20

ursed

1.19

ouple

1.18

rossover

1.18

ognitive

1.18

Activations Density 0.083%

phone calls and medical profession-related terms

No Comments

No Known Activations