INDEX

Explanations

phrases related to the concept of maintaining something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "maintained" and its variations

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

 heroes

-0.67

 Nobel

-0.65

 torch

-0.62

eto

-0.61

 dice

-0.58

 dinner

-0.56

 luck

-0.56

 inspiration

-0.56

eem

-0.56

ink

-0.56

POSITIVE LOGITS

 maintained

3.68

 maintains

2.10

 maintain

1.96

 retained

1.60

 maintaining

1.60

 kept

1.58

 preserved

1.55

 persisted

1.48

 operated

1.42

 managed

1.37

Activations Density 0.014%

phrases related to the concept of maintaining something

instances of the word "maintained" and its variations

No Comments

No Known Activations