INDEX

Explanations

locations or geographical entities, particularly in Southeast Asia

oai_token-act-pair · gpt-3.5-turbo

references to Malaysia and related geographic locations

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

 brut

-0.72

tops

-0.71

ptions

-0.65

velt

-0.64

ption

-0.63

ancies

-0.62

bern

-0.61

BAT

-0.61

 Farrell

-0.60

tower

-0.60

POSITIVE LOGITS

 Lumpur

1.83

 Airlines

1.39

 airliner

0.96

colm

0.89

 Kuala

0.88

 airlines

0.86

 Malays

0.84

0.81

 Malaysian

0.81

 Abdul

0.81

Activations Density 0.029%

locations or geographical entities, particularly in Southeast Asia

references to Malaysia and related geographic locations

No Comments

No Known Activations