INDEX

Explanations

phrases related to advancements or importance in a particular field

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

DragonMagazine

-1.01

imilar

-0.72

keys

-0.70

 Cosponsors

-0.70

items

-0.69

balls

-0.68

 sprayed

-0.67

Glass

-0.66

 borrowed

-0.65

Written

-0.65

POSITIVE LOGITS

 determining

1.27

 survival

1.19

 unlocking

1.17

 achieving

1.15

 sustaining

1.12

 maintaining

1.11

 understanding

1.09

 preserving

1.08

 ensuring

1.08

 shaping

1.00

Activations Density 0.143%

phrases related to advancements or importance in a particular field

No Comments

No Known Activations