INDEX

Explanations

words related to user interactions and generated content on websites

oai_token-act-pair · gpt-3.5-turbo

terms related to user-generated content and user experience

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

ĪĴ

-0.85

ounding

-0.79

 knots

-0.71

catentry

-0.70

rontal

-0.70

inventoryQuantity

-0.65

Reloaded

-0.65

 veins

-0.64

auld

-0.64

å¥³

-0.63

POSITIVE LOGITS

 Desire

0.76

 whom

0.71

EQ

0.71

 vide

0.68

 beware

0.68

Syd

0.67

defined

0.66

 waivers

0.65

 Jury

0.63

ipeg

0.62

Activations Density 0.501%

words related to user interactions and generated content on websites

terms related to user-generated content and user experience

No Comments

No Known Activations