INDEX

Explanations

phrases related to destruction or ruin

oai_token-act-pair · gpt-3.5-turbo

references to historical or physical destruction

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

leans

-0.86

soDeliveryDate

-0.74

 approving

-0.71

akeru

-0.71

pport

-0.71

assadors

-0.71

autions

-0.70

urai

-0.70

soType

-0.69

eret

-0.68

POSITIVE LOGITS

 carc

0.80

 habitat

0.78

 centrif

0.77

 Leviathan

0.77

 Fukushima

0.76

 fragments

0.75

 Annex

0.73

 seams

0.73

 habitats

0.73

 landfill

0.72

Activations Density 0.522%

phrases related to destruction or ruin

references to historical or physical destruction

No Comments

No Known Activations