INDEX

Explanations

content related to advertisements and potentially digital protection methods

oai_token-act-pair · gpt-3.5-turbo

indications of advertisements or promotional content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ħĭ

-0.73

 princ

-0.71

 internship

-0.67

 alignment

-0.64

 uniform

-0.64

 welding

-0.62

 faculties

-0.62

 dating

-0.62

Ń·

-0.61

 ausp

-0.59

POSITIVE LOGITS

Arcade

0.75

Anonymous

0.74

advertisement

0.73

Warning

0.71

VICE

0.70

*/

0.70

Correction

0.69

]"

0.69

ccording

0.69

Enlarge

0.68

Activations Density 0.057%

content related to advertisements and potentially digital protection methods

indications of advertisements or promotional content

No Comments

No Known Activations