INDEX

Explanations

words related to something that is not moving or changing, often used in a technological context

oai_token-act-pair · gpt-3.5-turbo

references to static states or 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

anth

-0.74

ittee

-0.73

lette

-0.72

rill

-0.72

nr

-0.71

nell

-0.71

ymes

-0.69

tta

-0.68

 alleges

-0.68

med

-0.67

POSITIVE LOGITS

 static

3.58

static

2.26

 Static

2.19

Static

2.10

 statically

1.83

 dynamic

1.48

Dynamic

1.36

 Dynamic

1.30

 stationary

1.30

 dynamically

1.27

Activations Density 0.018%

words related to something that is not moving or changing, often used in a technological context

references to static states or conditions

No Comments

No Known Activations