INDEX

Explanations

words related to electricity or power

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

terday

-0.81

itably

-0.75

 vain

-0.74

etheless

-0.73

ictions

-0.70

 sidel

-0.67

voc

-0.67

 routine

-0.67

 hepat

-0.67

matically

-0.66

POSITIVE LOGITS

 Room

1.14

 Point

1.11

Aid

1.10

Day

1.09

 Order

1.08

 Manager

1.07

Aid

1.07

Bank

1.07

 Builder

1.06

 Points

1.06

Activations Density 3.436%

words related to electricity or power

No Comments

No Known Activations