INDEX

Explanations

occurrences of the word "Grand" followed by a mentioning of a specific event or place

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "Grand Prix" or related events in a racing context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

yip

-0.93

PsyNetMessage

-0.79

lihood

-0.78

ued

-0.72

ople

-0.70

Netflix

-0.68

onies

-0.66

ause

-0.66

haar

-0.65

anwhile

-0.64

POSITIVE LOGITS

 Prix

1.48

 Theft

1.09

 Slam

1.08

master

1.04

 Finals

1.03

masters

0.97

finals

0.95

stand

0.95

cade

0.93

 Tour

0.87

Activations Density 0.008%

occurrences of the word "Grand" followed by a mentioning of a specific event or place

mentions of the term "Grand Prix" or related events in a racing context

No Comments

No Known Activations