INDEX

Explanations

content regarding tips and advice

oai_token-act-pair · gpt-3.5-turbo

references to advice or recommendations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ufact

-0.71

ruciating

-0.67

 LIMITED

-0.64

 Refugee

-0.63

 Annex

-0.63

 Asylum

-0.63

ords

-0.62

 Blaz

-0.61

 Accountability

-0.61

 Liberation

-0.60

POSITIVE LOGITS

 tips

1.55

tip

1.45

tip

1.37

Tip

1.29

Tips

1.29

tips

1.28

 Tips

1.27

Tip

1.05

 tipped

0.99

 tipping

0.97

Activations Density 0.011%

content regarding tips and advice

references to advice or recommendations

No Comments

No Known Activations