INDEX

Explanations

information related to technology tips and assistance

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

olina

-0.73

urst

-0.70

 inappropriately

-0.69

ELD

-0.68

ume

-0.65

amental

-0.65

interstitial

-0.63

 corrid

-0.63

orical

-0.62

ory

-0.62

POSITIVE LOGITS

 thankfully

0.86

 luckily

0.83

 survived

0.79

 sorted

0.77

thanks

0.76

 avoids

0.75

 speedy

0.74

 ðŁĻĤ

0.74

 calmed

0.73

 managed

0.73

Activations Density 0.255%

information related to technology tips and assistance

No Comments

No Known Activations