INDEX

Explanations

terms related to unethical or illegal activities, specifically corruption

oai_token-act-pair · gpt-3.5-turbo

instances and mentions of corruption

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

imus

-0.89

nee

-0.83

Â¯Â¯Â¯Â¯

-0.76

ofi

-0.75

AY

-0.73

amins

-0.72

lee

-0.71

estead

-0.70

emade

-0.70

ocard

-0.69

POSITIVE LOGITS

 corruption

1.11

 scandals

1.02

 corrupt

0.89

 Corruption

0.88

 corrupted

0.86

 scandal

0.86

 bribery

0.81

 disproportion

0.81

 penalty

0.79

 dealings

0.78

Activations Density 0.014%

terms related to unethical or illegal activities, specifically corruption

instances and mentions of corruption

No Comments

No Known Activations