INDEX

Explanations

the word "Grand" followed by a digit, specifically favoring "Grand 9" over "Grand 8"

oai_token-act-pair · gpt-3.5-turbo

references to the term "Grand" related to various locations or entities

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

ologies

-0.75

lihood

-0.74

ople

-0.73

qqa

-0.70

earance

-0.67

 smoker

-0.66

 numbered

-0.66

agen

-0.60

 livest

-0.60

 steroids

-0.60

POSITIVE LOGITS

mother

1.35

father

1.30

parents

1.27

 Prix

1.16

pa

1.14

 Rapids

1.11

 Theft

1.09

daughter

1.08

iosity

1.01

parent

0.99

Activations Density 0.019%

the word "Grand" followed by a digit, specifically favoring "Grand 9" over "Grand 8"

references to the term "Grand" related to various locations or entities

No Comments

No Known Activations

the word "Grand" followed by a digit, specifically favoring "Grand 9" over "Grand 8"

references to the term "Grand" related to various locations or entities

No Comments

No Known Activations