INDEX

Explanations

deeply held beliefs and values, particularly in the context of challenges to established beliefs or social norms

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

osponsors

-1.01

ateurs

-0.97

apps

-0.87

quer

-0.84

airs

-0.84

ilk

-0.82

paio

-0.82

paces

-0.82

©¶æ

-0.80

allows

-0.80

POSITIVE LOGITS

 belief

1.37

 tendency

1.28

 desire

1.27

 societal

1.25

 notion

1.20

 dissatisfaction

1.20

 misconception

1.19

 frustration

1.19

 animosity

1.18

 ideological

1.18

Activations Density 0.261%

deeply held beliefs and values, particularly in the context of challenges to established beliefs or social norms

No Comments

No Known Activations