INDEX

Explanations

mentions of destruction or damage

oai_token-act-pair · gpt-3.5-turbo

references to destruction or damage to entities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

soDeliveryDate

-0.94

soType

-0.83

akeru

-0.82

leans

-0.79

assadors

-0.76

tera

-0.76

say

-0.72

SPONSORED

-0.71

quickShipAvailable

-0.70

 approving

-0.69

POSITIVE LOGITS

 havoc

0.82

 Fukushima

0.82

 landfill

0.81

 carc

0.79

 limb

0.77

 seams

0.76

 uranium

0.76

 plutonium

0.74

 Leviathan

0.73

 habitat

0.72

Activations Density 0.433%

mentions of destruction or damage

references to destruction or damage to entities

No Comments

No Known Activations