INDEX

Explanations

phrases where someone has something in mind

oai_token-act-pair · gpt-3.5-turbo

references to concepts or ideas being considered or thought about

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

EStreamFrame

-0.75

AAA

-0.66

Professional

-0.63

rylic

-0.62

otta

-0.60

MpServer

-0.60

Growing

-0.58

 Mehran

-0.57

ittle

-0.57

TA

-0.57

POSITIVE LOGITS

fulness

0.83

lessly

0.81

ovie

0.74

eus

0.73

swer

0.72

lihood

0.72

 Kamp

0.71

lessness

0.69

uras

0.68

omsday

0.66

Activations Density 0.010%

phrases where someone has something in mind

references to concepts or ideas being considered or thought about

No Comments

No Known Activations