INDEX

Explanations

pronominal references indicating oneself

oai_token-act-pair · gpt-3.5-turbo

references to self-identity or self-ownership

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

ADVERTISEMENT

-0.62

ramid

-0.62

are

-0.55

airs

-0.51

 Commercial

-0.50

 Posted

-0.50

rx

-0.47

hak

-0.47

Filter

-0.46

oji

-0.45

POSITIVE LOGITS

 herself

4.42

 himself

2.51

 Himself

2.06

 oneself

2.03

 themselves

1.92

 myself

1.88

 ourselves

1.78

 itself

1.76

 yourself

1.75

 yourselves

1.65

Activations Density 0.010%

pronominal references indicating oneself

references to self-identity or self-ownership

No Comments

No Known Activations