INDEX

Explanations

keywords related to giving or receiving commands or instructions

oai_token-act-pair · gpt-3.5-turbo

references to processes involving orders

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

olkien

-0.69

 highlight

-0.65

ight

-0.65

azz

-0.63

itudinal

-0.63

FFER

-0.62

pton

-0.62

anger

-0.61

plug

-0.60

itch

-0.60

POSITIVE LOGITS

 orders

4.18

 Orders

3.11

orders

2.30

 order

2.26

 Order

1.95

 ORDER

1.91

Order

1.80

 ordering

1.74

 ordered

1.73

order

1.67

Activations Density 0.013%

keywords related to giving or receiving commands or instructions

references to processes involving orders

No Comments

No Known Activations