INDEX

Explanations

phrases indicating crucial importance or influence

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a significant role or contribution

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 heirs

-0.65

killed

-0.65

INGS

-0.64

soever

-0.62

LOG

-0.60

llor

-0.60

CVE

-0.59

GROUND

-0.59

pill

-0.58

 rappers

-0.57

POSITIVE LOGITS

 regards

1.25

roads

1.25

clus

1.15

effic

1.14

efficiency

1.12

 lieu

1.12

 vitro

1.08

 disguise

1.07

 conjunction

1.04

clusions

1.02

Activations Density 0.437%

phrases indicating crucial importance or influence

phrases indicating a significant role or contribution

No Comments

No Known Activations