INDEX

Explanations

references to the "Star Wars" franchise

oai_token-act-pair · gpt-3.5-turbo

references to the "Star Wars" franchise

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pd

-0.71

 pine

-0.69

mt

-0.67

 certify

-0.67

 commit

-0.64

eger

-0.63

 fingerprint

-0.62

 certified

-0.61

handle

-0.61

body

-0.61

POSITIVE LOGITS

 Wars

4.13

Wars

2.59

 wars

2.21

War

1.63

War

1.53

 Warfare

1.52

WAR

1.37

 Sith

1.32

 Trek

1.25

 Rebellion

1.25

Activations Density 0.018%

references to the "Star Wars" franchise

references to the "Star Wars" franchise

No Comments

No Known Activations