INDEX

Explanations

references to specific events or occurrences, particularly related to sports events or eclipses

oai_token-act-pair · gpt-3.5-turbo

references to weather events and astronomical phenomena

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

netflix

-0.75

SQ

-0.74

guyen

-0.74

 Flavoring

-0.71

Du

-0.68

 Geek

-0.67

 Lilith

-0.66

 affirmative

-0.66

 Survivors

-0.65

venant

-0.65

POSITIVE LOGITS

weather

2.71

 eclipse

1.48

 Watkins

1.45

 hiber

1.32

Weather

1.15

 groom

1.07

 Maxwell

1.01

yt

1.01

 herald

0.99

ussia

0.95

Activations Density 0.059%

references to specific events or occurrences, particularly related to sports events or eclipses

references to weather events and astronomical phenomena

No Comments

No Known Activations