INDEX

Explanations

elevations and altitudes mentioned in numerical form

oai_token-act-pair · gpt-3.5-turbo

references to elevation and altitude

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

å¥³

-0.85

soDeliveryDate

-0.83

ãģ®éŃĶ

-0.82

vous

-0.79

Cola

-0.77

Interstitial

-0.73

HEAD

-0.73

patch

-0.72

Â·Â·

-0.71

enegger

-0.70

POSITIVE LOGITS

 elevation

1.33

 altitude

1.22

 elev

1.01

 heights

0.98

ffen

0.90

 Elev

0.88

 clearance

0.84

 Everest

0.83

 sickness

0.81

 footh

0.80

Activations Density 0.008%

elevations and altitudes mentioned in numerical form

references to elevation and altitude

No Comments

No Known Activations