INDEX

Explanations

proper nouns and technical terms related to companies, locations, and products

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 hump

-0.78

 mechanically

-0.77

 unfavorable

-0.76

awa

-0.70

 seiz

-0.70

 typical

-0.70

 skilled

-0.70

 stride

-0.69

 blown

-0.69

inventoryQuantity

-0.68

POSITIVE LOGITS

Ltd

1.70

LLC

1.56

 Productions

1.35

LLP

1.34

Inc

1.31

 Limited

1.27

âĦ¢

1.27

 Solutions

1.25

LTD

1.25

 Studios

1.22

Activations Density 5.086%

proper nouns and technical terms related to companies, locations, and products

No Comments

No Known Activations