INDEX

Explanations

mentions of 'being' or variations of 'being' associated with evaluations or moral reflections

oai_token-act-pair · gpt-3.5-turbo

references to human beings and their characteristics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

DEP

-0.70

 tightening

-0.69

ICE

-0.68

 racks

-0.66

 tiss

-0.64

Effective

-0.64

 Operation

-0.63

Operation

-0.63

 Lans

-0.62

 supplemental

-0.62

POSITIVE LOGITS

 beings

0.95

plane

0.79

who

0.75

 inhab

0.75

puter

0.75

ĺħ

0.74

bender

0.73

 whom

0.71

 endowed

0.71

Unit

0.70

Activations Density 0.034%

mentions of 'being' or variations of 'being' associated with evaluations or moral reflections

references to human beings and their characteristics

No Comments

No Known Activations