INDEX

Explanations

instances of dishonest behavior, specifically cheating

oai_token-act-pair · gpt-3.5-turbo

terms related to dishonest behavior and cheating

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ŃĶ

-0.76

area

-0.71

escal

-0.70

areth

-0.69

oran

-0.67

Vert

-0.67

entin

-0.63

rez

-0.62

vig

-0.62

eric

-0.61

POSITIVE LOGITS

 cheating

0.88

 cheat

0.81

 cheated

0.80

ulence

0.78

 loophole

0.72

raud

0.71

ulent

0.70

 exploited

0.68

 loopholes

0.67

 wrongdoing

0.65

Activations Density 0.111%

instances of dishonest behavior, specifically cheating

terms related to dishonest behavior and cheating

No Comments

No Known Activations