INDEX

Explanations

celebrity names, food-related terms, and words relating to leadership and military actions

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Highlander

-0.69

WAYS

-0.69

press

-0.68

borg

-0.67

TERN

-0.67

metry

-0.65

xual

-0.64

 pregnant

-0.64

fill

-0.63

SHIP

-0.62

POSITIVE LOGITS

inelli

1.17

irtual

1.04

ies

1.00

oice

1.00

okes

0.97

ying

0.97

oked

0.95

iley

0.94

okers

0.94

ille

0.94

Activations Density 2.688%

celebrity names, food-related terms, and words relating to leadership and military actions

No Comments

No Known Activations