INDEX

Explanations

brand mentions or promotions for various products or services

oai_token-act-pair · gpt-3.5-turbo

instances of the "Sp" prefix, likely related to specific terms or brands

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

*/(

-0.82

 separatist

-0.71

xual

-0.68

 compuls

-0.67

sbm

-0.66

 mindful

-0.64

 Incarnation

-0.63

otype

-0.63

 fear

-0.63

 mind

-0.61

POSITIVE LOGITS

Sp

0.98

aghetti

0.89

irts

0.86

orp

0.85

iegel

0.82

onge

0.82

 Hels

0.82

ONSORED

0.81

itt

0.81

ools

0.80

Activations Density 0.006%

brand mentions or promotions for various products or services

instances of the "Sp" prefix, likely related to specific terms or brands

No Comments

No Known Activations