INDEX

Explanations

words related to mistreatment or mistreatment-like scenarios

oai_token-act-pair · gpt-3.5-turbo

references to varying degrees of mistreatment

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

jriwal

-0.81

thodox

-0.76

adesh

-0.75

lishes

-0.73

Books

-0.71

audi

-0.70

mobi

-0.66

Interstitial

-0.64

åĽ

-0.64

à¼

-0.64

POSITIVE LOGITS

reatment

1.28

reated

1.15

 mist

0.96

rans

0.94

ypes

0.91

reating

0.91

resses

0.88

ribut

0.82

mallow

0.81

arget

0.78

Activations Density 0.006%

words related to mistreatment or mistreatment-like scenarios

references to varying degrees of mistreatment

No Comments

No Known Activations