INDEX

Explanations

information about people and events according to sources

oai_token-act-pair · gpt-3.5-turbo

references to authoritative or credible sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

amacare

-0.88

à¼

-0.78

btn

-0.74

Magikarp

-0.74

 weaker

-0.72

jriwal

-0.71

blocking

-0.71

ðŁĳ

-0.71

Uber

-0.70

 ObamaCare

-0.70

POSITIVE LOGITS

 historian

1.72

 historians

1.43

 archives

1.43

 biography

1.42

 autobiography

1.31

 Archives

1.29

 Hist

1.24

 memoir

1.23

 curator

1.21

 Historical

1.20

Activations Density 0.533%

information about people and events according to sources

references to authoritative or credible sources

No Comments

No Known Activations