INDEX

Explanations

locations within buildings

oai_token-act-pair · gpt-3.5-turbo

locations or settings related to various scenarios and events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 occasions

-0.71

Americans

-0.69

 websites

-0.68

osate

-0.66

roots

-0.66

 Spears

-0.65

IDES

-0.64

 foes

-0.62

 verbs

-0.60

 entities

-0.59

POSITIVE LOGITS

 courtyard

1.45

 hallway

1.44

 room

1.33

 basement

1.28

 restroom

1.25

 attic

1.25

 lobby

1.23

 bathroom

1.22

 parking

1.20

 alley

1.18

Activations Density 0.352%

locations within buildings

locations or settings related to various scenarios and events

No Comments

No Known Activations