INDEX

Explanations

words related to physical or mental impairments and conditions

oai_token-act-pair · gpt-3.5-turbo

mentions of disability and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

furt

-0.86

soType

-0.83

hunt

-0.79

cold

-0.76

kov

-0.74

hens

-0.73

coat

-0.72

sis

-0.71

gres

-0.71

ulate

-0.70

POSITIVE LOGITS

 disabilities

0.92

 disability

0.86

 impair

0.82

abled

0.82

 Disability

0.81

 Rights

0.80

 rights

0.78

 suff

0.76

 Disabled

0.76

 inval

0.74

Activations Density 0.044%

words related to physical or mental impairments and conditions

mentions of disability and related terms

No Comments

No Known Activations