INDEX

Explanations

references to physical boundaries or restricted areas

oai_token-act-pair · gpt-3.5-turbo

terms related to boundaries and security measures

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

inn

-0.81

ttle

-0.78

å£«

-0.73

into

-0.71

ãĥīãĥ©

-0.69

inker

-0.68

andro

-0.68

umo

-0.67

Beer

-0.67

chin

-0.67

POSITIVE LOGITS

 perimeter

1.16

 fence

0.89

-+-+

0.85

 fences

0.79

 enclosure

0.74

 containment

0.74

 circumference

0.72

 fencing

0.71

 surrounds

0.71

 security

0.71

Activations Density 0.018%

references to physical boundaries or restricted areas

terms related to boundaries and security measures

No Comments

No Known Activations