INDEX

Explanations

mentions of specific geographical locations, especially those ending in 'istan'

oai_token-act-pair · gpt-3.5-turbo

references to the country "Balochistan" and its associated regions or related terms

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

upiter

-0.67

nard

-0.67

htaking

-0.66

ĨĴ

-0.65

hovah

-0.65

auld

-0.64

thora

-0.64

 gull

-0.63

angelo

-0.63

 payload

-0.63

POSITIVE LOGITS

istan

1.28

istani

0.95

uates

0.94

oglu

0.94

 Province

0.88

bia

0.84

 Tanzania

0.83

mA

0.78

uating

0.77

 Albania

0.76

Activations Density 0.022%

mentions of specific geographical locations, especially those ending in 'istan'

references to the country "Balochistan" and its associated regions or related terms

No Comments

No Known Activations