INDEX

Explanations

mentions of disabilities or actions related to disabilities

oai_token-act-pair · gpt-3.5-turbo

references to disability

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

 tide

-0.71

 anticipation

-0.68

Ã

-0.68

gie

-0.66

pour

-0.66

zin

-0.66

 mete

-0.65

allah

-0.64

 Wass

-0.64

 Guan

-0.63

POSITIVE LOGITS

 disabled

3.99

 Disabled

2.85

disabled

2.80

 disability

1.87

 disabling

1.74

 disable

1.73

 disabilities

1.70

 crippled

1.63

 Disability

1.56

 wheelchair

1.49

Activations Density 0.016%

mentions of disabilities or actions related to disabilities

references to disability

No Comments

No Known Activations