INDEX

Explanations

commands or coding instructions in structured text

oai_token-act-pair · gpt-3.5-turbo

sequences of characters or patterns in technical or specialized terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 earnest

-0.60

 practically

-0.59

 everyday

-0.57

 ÂŃ

-0.57

 Fundamental

-0.56

 empowering

-0.56

 past

-0.56

 accountability

-0.55

 behind

-0.55

 carrying

-0.55

POSITIVE LOGITS

isoft

1.21

odox

1.07

chwitz

1.03

duino

1.03

mble

1.02

redients

1.01

renheit

0.98

ertain

0.95

llular

0.93

iquette

0.91

Activations Density 0.126%

commands or coding instructions in structured text

sequences of characters or patterns in technical or specialized terms

No Comments

No Known Activations