INDEX

Explanations

words related to arrival or being in a specific location

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ifice

-0.71

akening

-0.67

rored

-0.67

 sake

-0.66

ult

-0.65

ality

-0.63

cription

-0.63

ister

-0.62

 Reserved

-0.62

urrent

-0.62

POSITIVE LOGITS

 shores

0.87

 doorstep

0.80

 Medina

0.78

eport

0.77

 Arri

0.73

 Lima

0.73

 Alam

0.72

 Athens

0.72

 Newark

0.72

 Albania

0.70

Activations Density 0.131%

words related to arrival or being in a specific location

No Comments

No Known Activations