INDEX

Explanations

mentions of military-related terms, particularly focusing on the word "Army" with different contexts

oai_token-act-pair · gpt-3.5-turbo

references to the Army

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

 refere

-0.70

 profiling

-0.70

metadata

-0.68

 transact

-0.67

 prof

-0.65

respons

-0.64

tty

-0.64

Ã°

-0.63

greg

-0.62

bnb

-0.61

POSITIVE LOGITS

 Army

3.68

Army

3.15

 army

2.30

 Navy

2.18

 Soldiers

1.72

 Armed

1.60

 Corps

1.59

 Naval

1.59

 Marines

1.58

Military

1.58

Activations Density 0.014%

mentions of military-related terms, particularly focusing on the word "Army" with different contexts

references to the Army

No Comments

No Known Activations