INDEX

Explanations

phrases related to moral principles or ethical standards

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

ittle

-0.68

geon

-0.67

sites

-0.67

sie

-0.65

fac

-0.65

availability

-0.65

azar

-0.65

ilant

-0.65

arf

-0.64

Raid

-0.63

POSITIVE LOGITS

 ideals

0.95

 principles

0.90

 embodied

0.87

 beliefs

0.83

 Values

0.82

 values

0.82

 creed

0.79

 diversity

0.78

 enshr

0.78

 tenets

0.76

Activations Density 0.049%

phrases related to moral principles or ethical standards

No Comments

No Known Activations