INDEX

Explanations

phrases related to authority figures or positions of power

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

wich

-0.78

avg

-0.77

ioned

-0.76

ilyn

-0.75

ewater

-0.70

het

-0.70

lda

-0.70

ipers

-0.69

 exting

-0.68

ilan

-0.67

POSITIVE LOGITS

 temptation

0.93

nor

0.63

 gorgeous

0.62

 Wildcats

0.61

 passion

0.61

 pleasures

0.60

 delight

0.59

rome

0.58

 showcasing

0.58

 provocation

0.57

Activations Density 0.268%

phrases related to authority figures or positions of power

No Comments

No Known Activations