INDEX

Explanations

mentions of the United States (U.S.) or related government institutions

oai_token-act-pair · gpt-3.5-turbo

references to the United States and its government actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

hots

-0.83

quickShipAvailable

-0.81

ÃŁ

-0.77

github

-0.76

Ô

-0.75

CHAPTER

-0.71

icester

-0.69

thodox

-0.69

LOD

-0.67

Pokemon

-0.66

POSITIVE LOGITS

 Constitution

0.95

 Citizenship

0.94

 Accountability

0.88

 Agency

0.81

 Ambassador

0.81

 ambassador

0.81

 Independence

0.80

 sovere

0.80

 Treasury

0.79

 electorate

0.76

Activations Density 0.269%

mentions of the United States (U.S.) or related government institutions

references to the United States and its government actions

No Comments

No Known Activations