INDEX

Explanations

references to ancient gods and mythical creatures, especially in a role-playing or fantasy context

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 Zucker

-0.88

espie

-0.85

displayText

-0.84

Case

-0.83

Article

-0.83

Hamilton

-0.82

Sports

-0.80

professional

-0.80

 endorsements

-0.80

Bloomberg

-0.79

POSITIVE LOGITS

 sorce

1.38

 sorcerer

1.31

 Norn

1.30

 Sorce

1.26

 undead

1.25

 mages

1.23

 Aether

1.22

 dragons

1.21

 Leviathan

1.21

 Undead

1.21

Activations Density 2.640%

references to ancient gods and mythical creatures, especially in a role-playing or fantasy context

No Comments

No Known Activations