INDEX

Explanations

punctuation usage and sentence endings

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

 credential

-0.83

intendent

-0.71

 elevated

-0.70

 authorized

-0.70

 illicit

-0.69

 cherished

-0.68

 startled

-0.67

 repud

-0.66

 divest

-0.65

 acute

-0.65

POSITIVE LOGITS

 Lastly

1.79

 Anyway

1.67

 Also

1.48

 Basically

1.47

 Hopefully

1.46

Anyway

1.44

 Additionally

1.41

Lastly

1.38

 Secondly

1.38

 Overall

1.34

Activations Density 0.381%

punctuation usage and sentence endings

No Comments

No Known Activations