INDEX

Explanations

names of programs, affiliations, and events related to academia and research

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

inventoryQuantity

-0.79

bir

-0.76

 skiing

-0.73

 slaughter

-0.72

paren

-0.70

panic

-0.69

depending

-0.69

 pawn

-0.68

ãĤ®

-0.68

 hiber

-0.68

POSITIVE LOGITS

 Manager

1.05

 Limited

1.02

Ltd

1.02

 Suite

1.00

Act

0.99

 Facility

0.98

 Scholarship

0.97

 Programme

0.95

 Fund

0.95

 Excellence

0.95

Activations Density 0.207%

names of programs, affiliations, and events related to academia and research

No Comments

No Known Activations