INDEX

Explanations

mentions of the death penalty, including both support and opposition

oai_token-act-pair · gpt-3.5-turbo

references to the death penalty

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

arity

-0.82

KY

-0.82

BN

-0.80

Remote

-0.79

å°Ĩ

-0.78

BIT

-0.77

é¾

-0.77

sonian

-0.77

Merit

-0.76

é¾įå

-0.75

POSITIVE LOGITS

 inmate

1.00

 inmates

0.88

 forfeiture

0.86

 incarcer

0.85

 torture

0.83

 manslaughter

0.82

 penalty

0.80

 overdose

0.79

 convict

0.78

 parole

0.76

Activations Density 0.050%

mentions of the death penalty, including both support and opposition

references to the death penalty

No Comments

No Known Activations