INDEX

Explanations

locations named "San"

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "San" in relation to various locations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

MEP

-0.76

 wing

-0.74

 balloon

-0.73

 lump

-0.71

 sheep

-0.71

 feathers

-0.70

 muscle

-0.70

 catapult

-0.70

 canvas

-0.68

ple

-0.68

POSITIVE LOGITS

San

3.74

SAN

2.24

San

2.20

san

2.05

Santa

1.84

SAN

1.79

Los

1.78

Sac

1.62

Miami

1.50

Oak

1.49

Activations Density 0.010%

locations named "San"

mentions of the word "San" in relation to various locations

No Comments

No Known Activations