INDEX

Explanations

instances where individuals or groups show a willingness to take action or make changes

oai_token-act-pair · gpt-3.5-turbo

references to willingness or unwillingness

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

og

-0.67

iner

-0.64

ogs

-0.63

ast

-0.62

uv

-0.60

-0.59

Mim

-0.59

ilk

-0.57

Neb

-0.57

quer

-0.57

POSITIVE LOGITS

 willingness

3.48

 unwillingness

2.36

 reluctance

1.89

 refusal

1.69

 readiness

1.69

 openness

1.68

 propensity

1.64

 willing

1.56

 inability

1.55

 desire

1.50

Activations Density 0.018%

instances where individuals or groups show a willingness to take action or make changes

references to willingness or unwillingness

No Comments

No Known Activations