INDEX

Explanations

locations, especially in the United States and Canada

oai_token-act-pair · gpt-3.5-turbo

mentions of U.S. states and cities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Frankfurt

-0.75

 Istanbul

-0.73

 Ankara

-0.71

emort

-0.66

 Karachi

-0.66

 Situation

-0.65

-0.63

 Podesta

-0.62

 Manila

-0.62

 roadside

-0.61

POSITIVE LOGITS

nesota

0.79

sylvania

0.78

fecture

0.68

union

0.67

jee

0.67

Ct

0.64

astern

0.64

Ļ

0.64

notation

0.63

 Hampshire

0.63

Activations Density 0.093%

locations, especially in the United States and Canada

mentions of U.S. states and cities

No Comments

No Known Activations