INDEX

Explanations

phrases related to justice or moral judgment

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

ĻĤ

-0.71

ason

-0.66

SPA

-0.65

 lapt

-0.63

 satell

-0.62

iaries

-0.61

 advoc

-0.60

worldly

-0.59

acas

-0.58

gard

-0.58

POSITIVE LOGITS

/"

1.15

 referring

0.95

["

0.84

 implying

0.79

0.75

 referencing

0.75

 refers

0.72

([

0.71

 meaning

0.71

("

0.70

Activations Density 0.655%

phrases related to justice or moral judgment

No Comments

No Known Activations