INDEX

Explanations

scientific terms and concepts related to research and studies

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

estate

-0.91

 whine

-0.80

 hates

-0.80

 hating

-0.76

iquette

-0.75

TRUMP

-0.74

 cursing

-0.72

 whore

-0.72

 lockout

-0.72

 cringe

-0.71

POSITIVE LOGITS

 researchers

1.53

 study

1.42

 computational

1.40

 experiments

1.39

 studies

1.37

Abstract

1.37

Researchers

1.36

Study

1.35

 analyses

1.34

 laboratory

1.34

Activations Density 6.729%

scientific terms and concepts related to research and studies

No Comments

No Known Activations