INDEX

Explanations

instructions or suggestions for modifying recipes and exploring creative combinations

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

 respectively

-0.60

 aftermath

-0.53

emale

-0.52

 allegedly

-0.52

 Mehran

-0.51

 Fundamental

-0.51

pires

-0.50

 roared

-0.50

 purported

-0.49

 Hussein

-0.48

POSITIVE LOGITS

 yourself

0.81

 manually

0.77

instead

0.76

 Yourself

0.71

 ASAP

0.71

 yourselves

0.70

 alternatives

0.69

morrow

0.66

agall

0.65

 preferably

0.64

Activations Density 0.659%

instructions or suggestions for modifying recipes and exploring creative combinations

No Comments

No Known Activations