INDEX

Explanations

word pairs showing a strong connection or relationship between two entities

oai_token-act-pair · gpt-3.5-turbo

phrases that suggest relationships or connections between entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Tru

-0.76

 Torment

-0.74

 Arri

-0.70

 Returning

-0.69

catentry

-0.68

anish

-0.66

 Mastery

-0.66

 Sett

-0.64

inventoryQuantity

-0.63

 Stamina

-0.62

POSITIVE LOGITS

unia

0.71

:[

0.69

pless

0.68

 differently

0.67

 deserve

0.65

 mean

0.64

 homework

0.63

 justice

0.63

 injustice

0.62

 grunt

0.62

Activations Density 0.531%

word pairs showing a strong connection or relationship between two entities

phrases that suggest relationships or connections between entities

No Comments

No Known Activations