INDEX

Explanations

mentions of watches, watch faces, and customization options related to watches

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

igl

-0.66

actionDate

-0.65

ussian

-0.62

NetMessage

-0.61

xon

-0.60

DragonMagazine

-0.60

luaj

-0.59

VIDIA

-0.57

hap

-0.57

 Flavoring

-0.57

POSITIVE LOGITS

Maker

0.92

maker

0.90

 maker

0.85

wright

0.83

builder

0.82

making

0.81

 builder

0.80

makers

0.77

holder

0.72

 supplier

0.70

Activations Density 3.710%

mentions of watches, watch faces, and customization options related to watches

No Comments

No Known Activations