INDEX

Explanations

locations in the United States, particularly focusing on cities and states

oai_token-act-pair · gpt-3.5-turbo

locations mentioned in the text

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

obin

-0.82

ULAR

-0.71

¦

-0.69

hift

-0.67

plot

-0.67

oton

-0.67

 extras

-0.66

ocial

-0.66

 companions

-0.65

SQL

-0.63

POSITIVE LOGITS

 respectively

0.87

 Dept

0.77

 Calif

0.73

 Strait

0.68

 southwest

0.68

etc

0.65

 Leban

0.65

accompanied

0.65

Ago

0.65

 northwest

0.64

Activations Density 0.134%

locations in the United States, particularly focusing on cities and states

locations mentioned in the text

No Comments

No Known Activations