INDEX

Explanations

financial and business terms, possibly related to fines, settlements, and company earnings

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Rebels

-0.78

 following

-0.68

 Conquest

-0.67

 Blizz

-0.66

 Nirvana

-0.65

 Hades

-0.64

 captcha

-0.64

 Korra

-0.63

 Extras

-0.63

 wide

-0.62

POSITIVE LOGITS

AB

1.33

CA

1.31

FP

1.30

SR

1.29

FI

1.25

SC

1.24

AG

1.22

FS

1.22

MA

1.22

AF

1.22

Activations Density 1.173%

financial and business terms, possibly related to fines, settlements, and company earnings

No Comments

No Known Activations