INDEX

Explanations

numeric values or indicators

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

 grop

-0.90

xual

-0.83

 tape

-0.80

 undecided

-0.79

 describ

-0.78

pse

-0.76

 possession

-0.75

 himself

-0.74

 causation

-0.73

 overth

-0.72

POSITIVE LOGITS

Learn

1.44

Join

1.43

Discover

1.40

Welcome

1.38

Features

1.36

Whether

1.34

Want

1.29

Explore

1.29

Our

1.28

Serv

1.26

Activations Density 0.220%

numeric values or indicators

No Comments

No Known Activations