INDEX

Explanations

proper names "Josh"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the name "Josh."

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

 subst

-0.87

holm

-0.69

alde

-0.67

Bridge

-0.66

castle

-0.65

sea

-0.64

arbon

-0.64

ront

-0.63

 carriage

-0.63

Ã©n

-0.62

POSITIVE LOGITS

 Josh

3.33

Josh

2.85

 Joshua

1.74

 Zach

1.51

Joshua

1.48

 Bryce

1.47

 Seth

1.45

 Greg

1.41

 Mitch

1.39

 Caleb

1.37

Activations Density 0.014%

proper names "Josh"

occurrences of the name "Josh."

No Comments

No Known Activations