INDEX

Explanations

phrases related to physical or metaphorical journeys

oai_token-act-pair · gpt-3.5-turbo

references to journeys or quests

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

 loopholes

-0.66

 Rowe

-0.62

 Marketable

-0.59

 elig

-0.59

 Treaty

-0.58

ãĥĲ

-0.58

 Auction

-0.57

 Entity

-0.56

 Availability

-0.56

email

-0.55

POSITIVE LOGITS

 toward

1.12

 towards

1.10

 Into

0.96

 into

0.95

 through

0.93

forward

0.92

 onward

0.92

 uphill

0.91

 journey

0.89

 downhill

0.89

Activations Density 0.184%

phrases related to physical or metaphorical journeys

references to journeys or quests

No Comments

No Known Activations