INDEX

Explanations

technical terms related to computer networks and information technology

oai_token-act-pair · gpt-3.5-turbo

mentions of lists and organized formats in written content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Incarn

-0.84

 Tsarnaev

-0.74

 Extrem

-0.74

 Superior

-0.72

 Fathers

-0.72

 Parenthood

-0.71

 Sons

-0.69

 Oriental

-0.69

 Mystic

-0.69

 indign

-0.68

POSITIVE LOGITS

map

1.02

code

1.02

links

1.01

ebin

1.01

 bookmark

0.99

elist

0.99

list

0.98

 google

0.97

wallet

0.97

ithub

0.96

Activations Density 0.484%

technical terms related to computer networks and information technology

mentions of lists and organized formats in written content

No Comments

No Known Activations