INDEX

Explanations

proper nouns and phrases related to news articles and events

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

opia

-0.73

angle

-0.72

ocracy

-0.69

senal

-0.68

 Himself

-0.68

ageddon

-0.67

ãĤ¹ãĥĪ

-0.67

BuyableInstoreAndOnline

-0.66

ãĤĭ

-0.64

uyomi

-0.63

POSITIVE LOGITS

 observers

1.04

 commenters

1.02

 commentators

1.01

 analysts

0.96

 scholars

0.93

 reviewers

0.85

 historians

0.84

 researchers

0.83

body

0.80

 people

0.79

Activations Density 2.483%

proper nouns and phrases related to news articles and events

No Comments

No Known Activations