INDEX

Explanations

words related to technical details and mechanisms, such as electronics and programming, and location-specific information like towns and sports teams

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

 Patriarch

-0.75

 fixation

-0.64

 Mechanics

-0.63

Dru

-0.63

 Madness

-0.62

 Breaker

-0.62

 Apostles

-0.59

 Nightmares

-0.59

 Divinity

-0.59

 Welch

-0.58

POSITIVE LOGITS

rican

1.38

riend

1.30

amily

1.28

rica

1.24

eatures

1.20

rost

1.20

raid

1.19

unction

1.17

ayette

1.16

rame

1.15

Activations Density 0.063%

words related to technical details and mechanisms, such as electronics and programming, and location-specific information like towns and sports teams

No Comments

No Known Activations