INDEX

Explanations

locations or places

oai_token-act-pair · gpt-3.5-turbo

references to geographic locations, particularly cities and towns

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

zona

-0.95

hner

-0.79

onz

-0.75

letal

-0.74

ollo

-0.73

inki

-0.72

Loading

-0.72

aundering

-0.71

ONY

-0.71

zes

-0.71

POSITIVE LOGITS

 Harbour

1.07

 Beach

0.98

 Harbor

0.93

 Spit

0.91

 waterfront

0.90

 Admir

0.90

 Yard

0.88

 Norfolk

0.83

 Roads

0.83

 beaches

0.83

Activations Density 0.033%

locations or places

references to geographic locations, particularly cities and towns

No Comments

No Known Activations