INDEX

Explanations

names, organizations, and locations mentioned in scholarly articles

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zona

-0.67

 Lumpur

-0.65

ĸļ

-0.63

ãĥ£

-0.62

 Miko

-0.61

wcs

-0.60

 Nadu

-0.60

 è£ıè

-0.59

 captcha

-0.58

ulhu

-0.56

POSITIVE LOGITS

ruary

1.03

owship

0.93

tein

0.73

ancial

0.72

heit

0.72

ascript

0.64

renheit

0.62

enhagen

0.61

iencies

0.60

 Furious

0.60

Activations Density 2.805%

names, organizations, and locations mentioned in scholarly articles

No Comments

No Known Activations