INDEX

Explanations

information related to projects, races, and endorsements in a political or competitive context

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

inventoryQuantity

-0.87

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.76

odynamics

-0.69

 Parables

-0.68

externalActionCode

-0.66

erity

-0.65

 Profit

-0.64

 Expect

-0.63

cffffcc

-0.62

 Citation

-0.62

POSITIVE LOGITS

 apiece

1.18

 besides

1.06

 including

1.02

 totaling

1.02

 vying

1.00

 imaginable

0.92

 consecut

0.89

 simultaneously

0.88

involved

0.87

 related

0.81

Activations Density 7.485%

information related to projects, races, and endorsements in a political or competitive context

No Comments

No Known Activations