INDEX

Explanations

terms related to political activities or movements that are opposing the government

oai_token-act-pair · gpt-3.5-turbo

references to government and related activities or sentiments

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.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

 Rath

-0.73

 Wilde

-0.71

 Peninsula

-0.70

east

-0.68

bleacher

-0.66

 Tome

-0.66

 Carson

-0.64

 Stam

-0.63

asts

-0.63

iera

-0.63

POSITIVE LOGITS

government

1.13

vernment

1.09

 sovere

0.84

etheless

0.83

rul

0.83

 administ

0.81

 shenan

0.80

 citiz

0.80

eco

0.80

 mathemat

0.79

Activations Density 0.006%

terms related to political activities or movements that are opposing the government

references to government and related activities or sentiments

No Comments

No Known Activations