INDEX

Explanations

emotional language and strong opinions

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

-0.63

utory

-0.62

 Nights

-0.59

ĨĴ

-0.59

 antid

-0.58

 ensures

-0.58

angan

-0.57

 Inqu

-0.57

 preventive

-0.56

 depends

-0.56

POSITIVE LOGITS

 unfold

1.17

 crumble

0.94

 firsthand

0.94

 emerge

0.85

 grow

0.80

 trending

0.75

 rise

0.75

 soar

0.74

 evolve

0.74

amorph

0.74

Activations Density 0.423%

emotional language and strong opinions

No Comments

No Known Activations