INDEX

Explanations

the word 'to'

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_fs6144-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_6144

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

6,144

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

quickShipAvailable

-0.68

 Closing

-0.66

 Caller

-0.66

listed

-0.65

 Needs

-0.64

soDeliveryDate

-0.63

 Warning

-0.60

ippi

-0.60

nil

-0.59

Choice

-0.59

POSITIVE LOGITS

 maximize

1.48

 ensure

1.48

 enhance

1.47

 create

1.45

 facilitate

1.43

 minimize

1.42

 achieve

1.39

 promote

1.38

 preserve

1.37

 determine

1.36

Activations Density 0.517%

the word 'to'

No Comments

No Known Activations