INDEX

Explanations

legal and political terms or descriptions related to government and legislation

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

DragonMagazine

-0.88

wic

-0.82

lication

-0.78

quished

-0.75

estern

-0.74

ãĥ¯ãĥ³

-0.67

âĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪ

-0.66

aram

-0.66

ledged

-0.66

arthed

-0.65

POSITIVE LOGITS

 thereby

1.22

 consequently

1.16

 hence

1.16

 prevents

1.14

 reduces

1.10

 undermines

1.09

 eliminates

1.07

 thus

1.07

 jeopard

1.06

 enhances

1.06

Activations Density 0.291%

legal and political terms or descriptions related to government and legislation

No Comments

No Known Activations