INDEX

Explanations

mentions of the word "Mongolia" or related terms

oai_token-act-pair · gpt-3.5-turbo

references to the Mongolian culture, people, or location

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

 cher

-0.74

rez

-0.74

Bet

-0.73

arget

-0.71

elle

-0.68

eries

-0.67

ere

-0.67

ulsion

-0.67

 Sirius

-0.66

Rem

-0.65

POSITIVE LOGITS

 Mong

3.44

 Mongol

2.83

 Mongolia

2.43

mong

1.85

ONG

1.60

jong

1.19

 Dmitry

1.14

 Cambod

1.09

ongs

1.06

osaurus

1.05

Activations Density 0.038%

mentions of the word "Mongolia" or related terms

references to the Mongolian culture, people, or location

No Comments

No Known Activations