INDEX

Explanations

the letter 'Z' followed by numbers

oai_token-act-pair · gpt-3.5-turbo

references to a specific individual or entity

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

 unpre

-0.68

 provoking

-0.62

Yel

-0.58

 Harm

-0.57

 life

-0.57

 minded

-0.57

 Sens

-0.56

selves

-0.54

not

-0.54

 Militia

-0.53

POSITIVE LOGITS

ombie

1.55

ombies

1.55

imbabwe

1.41

ERO

1.31

eus

1.27

odiac

1.26

ooming

1.23

immer

1.23

ebra

1.20

oom

1.19

Activations Density 0.020%

the letter 'Z' followed by numbers

references to a specific individual or entity

No Comments

No Known Activations