INDEX

Explanations

mentions of the location "Sydney"

oai_token-act-pair · gpt-3.5-turbo

mentions of the location Sydney

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

aired

-0.85

aid

-0.80

Magikarp

-0.72

ocobo

-0.71

OHN

-0.70

unct

-0.68

Torrent

-0.68

ever

-0.67

DonaldTrump

-0.67

arist

-0.66

POSITIVE LOGITS

 Sydney

1.05

 Harbour

1.01

 Lumpur

1.00

NSW

0.98

 Brisbane

0.98

 Adelaide

0.92

 Auckland

0.81

 Darling

0.80

NRL

0.79

 rall

0.79

Activations Density 0.005%

mentions of the location "Sydney"

mentions of the location Sydney

No Comments

No Known Activations