INDEX

Explanations

phrases or words related to veils or disguises

oai_token-act-pair · gpt-3.5-turbo

the word "ve" and its variations, indicating references to actions or states related to life or existence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

channelAvailability

-0.80

*/(

-0.79

 Cosponsors

-0.75

manship

-0.71

ONES

-0.69

INESS

-0.69

ORED

-0.68

DragonMagazine

-0.68

apter

-0.66

å§«

-0.66

POSITIVE LOGITS

ggie

1.04

ices

1.04

uve

0.88

mber

0.87

cks

0.85

uer

0.84

lez

0.84

gging

0.83

zx

0.82

xt

0.81

Activations Density 0.005%

phrases or words related to veils or disguises

the word "ve" and its variations, indicating references to actions or states related to life or existence

No Comments

No Known Activations