INDEX

Explanations

phrases related to threatening or demanding actions

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

 Passage

-0.65

COMPLE

-0.64

AMP

-0.64

iculty

-0.59

VERTISEMENT

-0.58

 purported

-0.58

ELD

-0.57

cler

-0.55

MET

-0.54

ournal

-0.53

POSITIVE LOGITS

 yourselves

1.52

 yourself

1.19

 Yourself

0.96

thy

0.92

ya

0.86

 your

0.86

esty

0.85

 kidding

0.85

 thou

0.83

ye

0.81

Activations Density 0.217%

phrases related to threatening or demanding actions

No Comments

No Known Activations