INDEX

Explanations

mentions of the word "pilates" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to "Pil" or words related to Pilates

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

anwhile

-0.92

 Carbuncle

-0.85

 EntityItem

-0.79

 Dwell

-0.68

Byr

-0.68

 Baird

-0.66

upon

-0.65

ires

-0.64

 Debor

-0.64

 Clapper

-0.63

POSITIVE LOGITS

oted

1.00

crow

0.97

wings

0.96

ando

0.96

icate

0.90

cipled

0.89

chard

0.86

ots

0.84

cone

0.83

awar

0.83

Activations Density 0.019%

mentions of the word "pilates" in various contexts

references to "Pil" or words related to Pilates

No Comments

No Known Activations