INDEX

Explanations

proper nouns or names related to identification

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the subject or agent associated with an action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 resil

-0.76

ETF

-0.71

itto

-0.71

 countered

-0.66

 destined

-0.65

ettes

-0.63

 recre

-0.63

edo

-0.62

ettle

-0.62

eal

-0.62

POSITIVE LOGITS

 virtue

0.95

laws

0.93

pass

0.80

 Seymour

0.73

 Forbes

0.70

ä½¿

0.68

Ĭ±

0.68

products

0.68

law

0.67

 statute

0.66

Activations Density 0.133%

proper nouns or names related to identification

phrases indicating the subject or agent associated with an action

No Comments

No Known Activations