INDEX

Explanations

references to actions taken by or concerning a specific individual

oai_token-act-pair · gpt-3.5-turbo

instances of the word "himself."

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

ramid

-0.67

ADVERTISEMENT

-0.59

updated

-0.57

angular

-0.57

are

-0.57

 âĢº

-0.56

mary

-0.54

prov

-0.53

ymm

-0.52

Reporting

-0.51

POSITIVE LOGITS

 himself

4.03

 Himself

3.10

 herself

2.98

 oneself

2.22

 themselves

2.15

 myself

1.99

 yourself

1.91

 ourselves

1.90

 itself

1.87

 yourselves

1.71

Activations Density 0.016%

references to actions taken by or concerning a specific individual

instances of the word "himself."

No Comments

No Known Activations