INDEX

Explanations

medical conditions or health-related terms

oai_token-act-pair · gpt-3.5-turbo

mentions of medical or physical conditions

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

hots

-0.86

apers

-0.74

angel

-0.69

hers

-0.68

usercontent

-0.67

igr

-0.67

ucks

-0.65

ango

-0.63

rations

-0.63

hey

-0.63

POSITIVE LOGITS

 condition

4.06

 Condition

2.79

condition

2.76

Condition

2.32

 conditions

2.25

 Conditions

1.99

 conditioning

1.64

 conditioned

1.50

 circumstance

1.28

 situation

1.28

Activations Density 0.019%

medical conditions or health-related terms

mentions of medical or physical conditions

No Comments

No Known Activations