INDEX

Explanations

mentions of the word "Kids"

oai_token-act-pair · gpt-3.5-turbo

references to children and content related to them

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 bombard

-0.75

 prec

-0.73

 explosive

-0.70

 polyg

-0.69

 circumstance

-0.69

 abund

-0.68

 guarant

-0.67

 understatement

-0.65

 opera

-0.65

ecd

-0.65

POSITIVE LOGITS

 Kids

3.03

Kids

1.80

kids

1.14

 Youth

1.13

Kid

1.11

 Idea

1.09

 Shield

1.08

 Shot

1.02

 Guys

1.01

 Talk

1.01

Activations Density 0.019%

mentions of the word "Kids"

references to children and content related to them

No Comments

No Known Activations