INDEX

Explanations

phrases emphasizing contrast or comparison between different entities

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

coni

-0.53

ãĤ´ãĥ³

-0.52

abee

-0.51

 Pipeline

-0.50

¬¼

-0.49

atern

-0.49

 Confederation

-0.49

èĢħ

-0.48

chell

-0.48

 corrid

-0.47

POSITIVE LOGITS

huh

1.10

 please

0.94

etc

0.87

but

0.84

 preferably

0.83

eh

0.83

tho

0.83

please

0.82

 haha

0.80

 yeah

0.79

Activations Density 0.338%

phrases emphasizing contrast or comparison between different entities

No Comments

No Known Activations