INDEX

Explanations

phrases related to improving something, such as clarity or grip

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

ãĥ¯ãĥ³

-0.69

Wer

-0.64

 Nobel

-0.63

 earnest

-0.61

 Passion

-0.61

 Soccer

-0.61

 Pelosi

-0.60

 born

-0.60

 Waterloo

-0.60

wife

-0.59

POSITIVE LOGITS

 minimize

1.17

 maximize

1.16

 accommodate

1.07

 avoid

1.05

avoid

1.02

 cumbers

1.00

 facilitate

0.99

 maximizing

0.95

 insure

0.95

 ensure

0.94

Activations Density 0.511%

phrases related to improving something, such as clarity or grip

No Comments

No Known Activations