INDEX

Explanations

mentions of the term "Pirate"

oai_token-act-pair · gpt-3.5-turbo

references to "Pirate" and "Pirates"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

enegger

-0.84

ires

-0.76

uating

-0.76

itsch

-0.76

pres

-0.75

andra

-0.73

eger

-0.73

gow

-0.72

trak

-0.71

upon

-0.71

POSITIVE LOGITS

Pir

1.09

 Pirate

1.08

 Pirates

0.98

 pirates

0.96

 Luffy

0.86

 Penguin

0.82

Pir

0.79

 pirate

0.79

 Ship

0.78

 Parties

0.78

Activations Density 0.006%

mentions of the term "Pirate"

references to "Pirate" and "Pirates"

No Comments

No Known Activations