INDEX

Explanations

phrases relating to legal or emergency alerts

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

icz

-0.90

 fermented

-0.80

 quir

-0.79

iffe

-0.78

 nodd

-0.76

 confid

-0.75

 freight

-0.74

 gaming

-0.73

 induct

-0.73

 forwarding

-0.72

POSITIVE LOGITS

 Includes

1.40

 Also

1.36

 Especially

1.31

 Retrieved

1.29

 Appears

1.27

 Seems

1.25

jpg

1.20

 Contains

1.19

 Allows

1.17

 Doesn

1.17

Activations Density 1.608%

phrases relating to legal or emergency alerts

No Comments

No Known Activations