INDEX

Explanations

mentions of the word "perm" or its variations

oai_token-act-pair · gpt-3.5-turbo

references to the concept of permanence

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

æĸ

-0.77

Eye

-0.74

 Scarlet

-0.71

 Dayton

-0.67

aughty

-0.64

service

-0.63

 Bian

-0.63

NCT

-0.63

joice

-0.62

lings

-0.62

POSITIVE LOGITS

 perm

4.27

perm

2.30

 poss

1.05

acan

0.90

 polar

0.85

 combinations

0.84

 melting

0.84

 subsequ

0.83

 pard

0.83

mul

0.82

Activations Density 0.023%

mentions of the word "perm" or its variations

references to the concept of permanence

No Comments

No Known Activations