INDEX

Explanations

phrases related to expressing a strong opinion or belief

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uyomi

-0.71

NetMessage

-0.65

ername

-0.64

 reperto

-0.62

 pronounce

-0.60

 inherit

-0.59

beam

-0.58

ynes

-0.58

zers

-0.57

foundation

-0.57

POSITIVE LOGITS

 sorts

1.47

 course

1.28

 ours

1.09

 theirs

1.06

icial

1.01

course

1.00

 yours

0.95

 hers

0.94

ortunately

0.83

 Excellence

0.82

Activations Density 0.854%

phrases related to expressing a strong opinion or belief

No Comments

No Known Activations