INDEX

Explanations

words related to names of people

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter 'i' in various contexts or patterns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 advertisers

-0.76

 overwhelming

-0.67

 necessities

-0.65

 shareholders

-0.65

 loopholes

-0.64

 innocence

-0.64

 millions

-0.64

 depreciation

-0.64

 ones

-0.62

 hospitality

-0.62

POSITIVE LOGITS

oglu

1.55

uchi

1.55

ishi

1.42

eva

1.40

cki

1.37

onen

1.28

oka

1.28

emi

1.26

aga

1.26

aki

1.24

Activations Density 0.163%

words related to names of people

occurrences of the letter 'i' in various contexts or patterns

No Comments

No Known Activations