INDEX

Explanations

words related to news articles, interviews, and information

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ń·

-0.71

illac

-0.70

Hoo

-0.61

 Blues

-0.61

 Chero

-0.60

 Reloaded

-0.59

Ĥª

-0.58

 Butterfly

-0.58

Kuh

-0.58

 Sorce

-0.57

POSITIVE LOGITS

depth

1.14

advert

1.04

built

1.03

line

0.99

sight

0.99

browser

0.99

development

0.98

between

0.96

camera

0.94

game

0.92

Activations Density 0.016%

words related to news articles, interviews, and information

No Comments

No Known Activations