INDEX

Explanations

ways or methods to achieve something or solve a problem

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

cour

-0.77

Requ

-0.74

assisted

-0.74

Alert

-0.72

idium

-0.69

angering

-0.67

imony

-0.65

warn

-0.64

 cautioned

-0.64

upt

-0.64

POSITIVE LOGITS

 foothold

1.05

 suitable

1.03

 ways

0.97

 solution

0.95

 groove

0.94

 scapego

0.94

 loophole

0.94

 balance

0.85

 receptive

0.85

 equilibrium

0.84

Activations Density 0.138%

ways or methods to achieve something or solve a problem

No Comments

No Known Activations