INDEX

Explanations

titles or names related to a specific historical figure or location

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

NetMessage

-0.78

 Flavoring

-0.75

PDATE

-0.71

hift

-0.70

++++++++++++++++

-0.66

ictions

-0.66

ressing

-0.66

IAL

-0.66

ruct

-0.66

Org

-0.65

POSITIVE LOGITS

shire

1.69

hurst

0.88

 Memorial

0.85

dale

0.81

 Parish

0.81

 Park

0.79

incoln

0.78

 Lincoln

0.77

 Heights

0.76

 Cathedral

0.75

Activations Density 0.017%

titles or names related to a specific historical figure or location

No Comments

No Known Activations