INDEX

Explanations

terms related to computer networks or organizations

oai_token-act-pair · gpt-3.5-turbo

mentions of various types of networks

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.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

 Dough

-0.85

itars

-0.75

shall

-0.70

 Dragonbound

-0.68

Zin

-0.67

 Asuka

-0.65

Tit

-0.65

aurus

-0.64

 Sailor

-0.64

 Summers

-0.63

POSITIVE LOGITS

 network

3.64

 networks

2.93

network

2.63

 Network

2.50

Network

2.43

 Networks

2.16

 NETWORK

2.08

 networking

1.99

 channel

1.38

 system

1.31

Activations Density 0.025%

terms related to computer networks or organizations

mentions of various types of networks

No Comments

No Known Activations