INDEX

Explanations

words related to reconstruction or repurposing of items or processes

oai_token-act-pair · gpt-3.5-turbo

prefixes and terms related to modification or alteration

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 willingness

-0.74

 sincerity

-0.72

 seriousness

-0.70

 gamble

-0.69

 belief

-0.68

lihood

-0.67

 inability

-0.67

 desire

-0.65

 beliefs

-0.65

 attitude

-0.65

POSITIVE LOGITS

urated

1.25

alysed

1.24

idered

1.21

ised

1.19

idden

1.19

handled

1.18

itored

1.17

aded

1.17

rolled

1.16

ased

1.16

Activations Density 0.273%

words related to reconstruction or repurposing of items or processes

prefixes and terms related to modification or alteration

No Comments

No Known Activations