INDEX

Explanations

names associated with military or governmental figures

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

 specificity

-0.67

iaries

-0.66

irlf

-0.65

okemon

-0.65

arthy

-0.64

orsi

-0.63

iless

-0.61

icable

-0.60

oun

-0.58

 Carbuncle

-0.58

POSITIVE LOGITS

utenant

1.53

berman

1.16

uten

1.10

ge

1.02

ÃŁ

0.97

bling

0.93

 detector

0.91

pard

0.91

hart

0.91

Lie

0.87

Activations Density 0.030%

names associated with military or governmental figures

No Comments

No Known Activations