INDEX

Explanations

the word "to" in sentences where it expresses negative consequences or criticism

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

listed

-0.73

quickShipAvailable

-0.72

points

-0.70

soDeliveryDate

-0.68

pins

-0.67

soever

-0.66

calling

-0.65

sle

-0.63

Dur

-0.63

felt

-0.62

POSITIVE LOGITS

 introduce

1.32

 educate

1.30

 observe

1.26

 celebrate

1.25

 capitalize

1.25

 declare

1.25

 create

1.24

 revisit

1.24

 engage

1.23

 invoke

1.22

Activations Density 1.087%

the word "to" in sentences where it expresses negative consequences or criticism

No Comments

No Known Activations