INDEX

Explanations

words related to tension or conflict

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

Negative Logits

hyde

-0.79

SHIP

-0.77

merce

-0.75

culosis

-0.73

STON

-0.66

BALL

-0.62

wolves

-0.62

pei

-0.62

å§«

-0.62

ventory

-0.61

POSITIVE LOGITS

angle

1.20

angled

1.20

angling

1.18

aining

1.15

aditional

1.11

atum

1.11

addle

1.07

ung

1.06

ained

1.06

aying

1.05

Activations Density 0.011%

words related to tension or conflict

No Known Activations