INDEX

Explanations

terms related to software installation and activation processes

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ewitness

-0.79

bold

-0.68

enegger

-0.68

 Curry

-0.65

inately

-0.63

iverse

-0.63

hest

-0.62

arag

-0.62

uate

-0.62

overe

-0.61

POSITIVE LOGITS

 techniques

1.12

 methods

1.07

 procedures

1.05

 process

1.04

 mechanisms

1.02

 mechanism

1.00

 capability

1.00

 efforts

0.97

eering

0.94

 capabilities

0.94

Activations Density 4.695%

terms related to software installation and activation processes

No Comments

No Known Activations