INDEX

Explanations

quotations attributed to someone with varying emotional tones

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

ersen

-0.74

rone

-0.68

arlane

-0.65

OTAL

-0.64

 answ

-0.63

iers

-0.63

iership

-0.63

apons

-0.62

asio

-0.60

Side

-0.60

POSITIVE LOGITS

 follows

1.10

 "...

0.98

 "âĢ¦

0.97

0.93

"[

0.90

"'

0.84

pires

0.82

 well

0.81

pired

0.81

 having

0.81

Activations Density 0.061%

quotations attributed to someone with varying emotional tones

No Comments

No Known Activations