INDEX

Explanations

names of specific locations, especially cities

oai_token-act-pair · gpt-3.5-turbo

location names, specifically cities

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

arial

-0.89

vana

-0.88

egal

-0.73

hart

-0.72

otic

-0.70

idel

-0.70

arat

-0.68

ird

-0.68

vich

-0.68

icles

-0.67

POSITIVE LOGITS

shire

0.84

 County

0.81

 Jail

0.77

 Gazette

0.75

 mayoral

0.75

 skyline

0.74

 Municip

0.69

 Niagara

0.68

 Raleigh

0.68

incinn

0.67

Activations Density 0.032%

names of specific locations, especially cities

location names, specifically cities

No Comments

No Known Activations