INDEX

Explanations

phrases related to residential buildings, decision-making, psychological terms related to assertiveness and passivity, Google Analytics terms, programming languages like Lisp, and phrases related to admin tools in an online platform

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

github

-0.69

hig

-0.68

sych

-0.64

rehens

-0.63

Enlarge

-0.62

atom

-0.61

meier

-0.60

ãĥ©ãĥ³

-0.60

ãĥ£

-0.60

POSITIVE LOGITS

 itself

0.82

 minus

0.80

 herself

0.71

rul

0.70

 equivalents

0.68

 altogether

0.68

 himself

0.66

 anyway

0.65

 plus

0.64

 lovers

0.63

Activations Density 9.001%

phrases related to residential buildings, decision-making, psychological terms related to assertiveness and passivity, Google Analytics terms, programming languages like Lisp, and phrases related to admin tools in an online platform

No Comments

No Known Activations