INDEX

Explanations

mentions or references to wives with varying levels of activation

oai_token-act-pair · gpt-3.5-turbo

mentions of a spouse or partner

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

kefeller

-0.97

osta

-0.80

vernment

-0.75

ostic

-0.73

umbn

-0.72

emonium

-0.70

 ceilings

-0.69

pmwiki

-0.68

 cryptoc

-0.68

aneously

-0.68

POSITIVE LOGITS

wife

1.11

 Michelle

0.96

 Kate

0.95

 Melania

0.94

 Martha

0.93

 Anne

0.91

 Sarah

0.89

 Julia

0.88

 Catherine

0.87

 Esther

0.86

Activations Density 0.041%

mentions or references to wives with varying levels of activation

mentions of a spouse or partner

No Comments

No Known Activations