INDEX

Explanations

instances of disregarding rules or consequences

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

arta

-0.83

wang

-0.81

agues

-0.78

stem

-0.75

english

-0.71

outer

-0.71

une

-0.71

wake

-0.70

link

-0.68

 Ahead

-0.67

POSITIVE LOGITS

 murder

1.04

 murdering

0.95

 manslaughter

0.92

 Murder

0.90

 mischief

0.87

 impunity

0.85

 reckless

0.84

 violating

0.83

 exploiting

0.80

 immoral

0.79

Activations Density 0.052%

instances of disregarding rules or consequences

No Comments

No Known Activations