INDEX

Explanations

mentions of political figures or positions, particularly focusing on the word "minister."

oai_token-act-pair · gpt-3.5-turbo

mentions of government officials or ministers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

hots

-0.68

Oxy

-0.63

Pit

-0.62

 Trace

-0.59

{"

-0.58

POS

-0.58

eca

-0.58

android

-0.58

âľ

-0.57

âĸ

-0.57

POSITIVE LOGITS

 minister

3.76

 Minister

2.75

 ministers

2.54

 ministry

2.44

 Ministers

2.09

 ministries

2.02

 ministerial

1.86

 Ministry

1.78

 secretary

1.62

 pastor

1.53

Activations Density 0.012%

mentions of political figures or positions, particularly focusing on the word "minister."

mentions of government officials or ministers

No Comments

No Known Activations