INDEX

Explanations

political and governmental terms and entities

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

maxwell

-0.79

 Twain

-0.79

 chops

-0.77

 crates

-0.73

Jones

-0.71

 crawl

-0.71

thy

-0.70

 dinosaurs

-0.69

 wallets

-0.69

 rive

-0.69

POSITIVE LOGITS

 Organisation

1.24

 Programme

1.16

 Directorate

1.16

 Assembly

1.08

 Defence

1.03

odo

1.02

 Council

1.00

 Tribunal

0.99

 Centre

0.97

 Prosecutor

0.96

Activations Density 0.187%

political and governmental terms and entities

No Comments

No Known Activations