INDEX

Explanations

mention of legal terms related to criminal offenses

oai_token-act-pair · gpt-3.5-turbo

terms associated with legal violations or criminal activities

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

bors

-0.75

 modeling

-0.71

DOD

-0.67

 FEMA

-0.66

 NASA

-0.66

spe

-0.66

 stabilized

-0.65

pled

-0.63

--

-0.63

 rumored

-0.62

POSITIVE LOGITS

 offence

4.32

 offences

2.66

 offense

2.33

 defence

1.60

 offenses

1.46

 Offense

1.39

 humour

1.38

 offender

1.36

 defences

1.36

 licence

1.32

Activations Density 0.008%

mention of legal terms related to criminal offenses

terms associated with legal violations or criminal activities

No Comments

No Known Activations