INDEX

Explanations

addresses

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Analy

-0.66

anything

-0.65

 causation

-0.65

limits

-0.65

Unknown

-0.63

agging

-0.62

CVE

-0.59

iris

-0.58

 excellence

-0.58

 freezing

-0.58

POSITIVE LOGITS

 Blvd

1.03

Ave

0.92

Rd

0.88

illion

0.83

+(

0.79

 acres

0.78

 apiece

0.75

âĦ¢:

0.67

IRO

0.65

RD

0.65

Activations Density 0.345%

addresses

No Comments

No Known Activations