INDEX

Explanations

mentions of physical locations or directions

oai_token-act-pair · gpt-3.5-turbo

keywords related to movement and activity

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

forcer

-0.70

istor

-0.69

timer

-0.63

iece

-0.62

sequence

-0.62

iom

-0.61

utan

-0.61

outine

-0.60

raper

-0.59

unit

-0.59

POSITIVE LOGITS

 ones

1.46

nesses

1.10

mails

1.05

rows

1.03

 fronts

1.02

aries

1.01

 positions

0.99

 outputs

0.99

 votes

0.99

 voices

0.99

Activations Density 0.602%

mentions of physical locations or directions

keywords related to movement and activity

No Comments

No Known Activations