INDEX

Explanations

proper nouns related to political figures

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

exempt

-0.83

road

-0.78

nir

-0.78

rl

-0.77

agra

-0.76

ahime

-0.76

ebook

-0.74

yk

-0.74

nexus

-0.74

rf

-0.74

POSITIVE LOGITS

 Rodham

1.07

 Trump

1.07

 Reagan

1.03

 Wallace

0.99

 Trudeau

0.99

 Clinton

0.98

 Johnson

0.98

 Gates

0.98

 Kennedy

0.97

 Obama

0.95

Activations Density 1.202%

proper nouns related to political figures

No Comments

No Known Activations