INDEX

Explanations

words related to different aspects of psychology, such as psychological status, personality traits, emotional intelligence, psychological distress, and psychological warfare

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

tower

-1.06

lain

-0.81

lessly

-0.79

bare

-0.79

plin

-0.78

urat

-0.77

ered

-0.75

upon

-0.74

alos

-0.74

ortium

-0.74

POSITIVE LOGITS

ized

1.30

izing

1.25

ization

1.24

ities

1.20

ity

1.17

ised

1.14

ising

1.12

 significance

1.11

isation

1.08

izes

1.04

Activations Density 3.690%

words related to different aspects of psychology, such as psychological status, personality traits, emotional intelligence, psychological distress, and psychological warfare

No Comments

No Known Activations