INDEX

Explanations

mentions of cheating or deceptive behavior

oai_token-act-pair · gpt-3.5-turbo

words related to dishonest behavior, particularly cheating

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

oran

-0.70

oan

-0.67

 Archdemon

-0.66

DragonMagazine

-0.63

Pain

-0.62

istan

-0.62

RIS

-0.60

Oper

-0.60

affer

-0.59

english

-0.58

POSITIVE LOGITS

 cheat

0.91

 cheating

0.83

ulative

0.82

herer

0.79

 cheated

0.76

cheat

0.73

yre

0.69

chet

0.67

 sheet

0.66

 Bastard

0.64

Activations Density 0.012%

mentions of cheating or deceptive behavior

words related to dishonest behavior, particularly cheating

No Comments

No Known Activations