INDEX

Explanations

questions and phrases that indicate curiosity or inquiry

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

estyles

-0.89

ursions

-0.83

 overfl

-0.80

 deployments

-0.76

 refin

-0.72

 withdrawals

-0.72

ulations

-0.71

 offsets

-0.71

 registrations

-0.71

 tours

-0.70

POSITIVE LOGITS

 whoever

1.07

 Yourself

0.96

 himself

0.89

 Someone

0.86

 Adolf

0.83

 myself

0.82

 Himself

0.82

omever

0.81

 somebody

0.80

 someone

0.80

Activations Density 0.206%

questions and phrases that indicate curiosity or inquiry

No Comments

No Known Activations