INDEX

Explanations

words related to moral values and philosophical concepts, particularly about beliefs, virtues, and ethical dilemmas

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

ovember

-0.75

senal

-0.69

 Reported

-0.67

recomm

-0.65

eligible

-0.63

 Rounds

-0.63

HB

-0.62

osponsors

-0.62

 billed

-0.62

 Reports

-0.60

POSITIVE LOGITS

 consciousness

1.08

 oneself

1.06

 humanity

1.04

 truths

1.03

 transcend

1.01

 righteousness

1.01

 ideals

0.98

 intellect

0.98

 individuality

0.95

 pleasures

0.93

Activations Density 5.268%

words related to moral values and philosophical concepts, particularly about beliefs, virtues, and ethical dilemmas

No Comments

No Known Activations