INDEX

Explanations

proper nouns or names of places

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

Not in Any Lists

No Comments

Negative Logits

 Monstrous

-0.67

 phyl

-0.65

 explan

-0.64

âĸ¬âĸ¬

-0.63

 defic

-0.59

 interf

-0.58

IGHTS

-0.57

 cigars

-0.56

 omission

-0.56

ccording

-0.56

POSITIVE LOGITS

olitan

1.43

oleon

1.42

ropri

1.21

ocalypse

1.20

ocalyptic

1.16

rison

1.06

osite

1.06

olitics

1.05

edia

1.05

onent

1.05

Activations Density 0.034%

proper nouns or names of places

No Comments

No Known Activations