INDEX

Explanations

phrases related to the number "two."

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

yles

-0.69

 Opportun

-0.67

 Challenges

-0.63

 Caption

-0.62

 Mysteries

-0.60

 embassies

-0.59

Rac

-0.57

 Secrets

-0.56

 Indust

-0.56

 Syndicate

-0.55

POSITIVE LOGITS

fold

1.24

teenth

1.14

 dimensional

1.06

peat

1.02

quel

0.99

dimensional

0.97

0.95

âĢĳ

0.90

some

0.88

eenth

0.83

Activations Density 0.058%

phrases related to the number "two."

No Comments

No Known Activations