INDEX

Explanations

phrases instructing to share or email content

oai_token-act-pair · gpt-3.5-turbo

references to the word "this."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

aneers

-0.69

alties

-0.69

Ø§

-0.68

owers

-0.68

cases

-0.67

Examples

-0.66

 roots

-0.64

ono

-0.63

 ceilings

-0.63

RESULTS

-0.62

POSITIVE LOGITS

 article

1.10

 item

0.94

 slideshow

0.84

 page

0.83

 ARTICLE

0.83

 document

0.82

 site

0.81

 webpage

0.79

article

0.78

 essay

0.75

Activations Density 0.075%

phrases instructing to share or email content

references to the word "this."

No Comments

No Known Activations