INDEX

Explanations

names related to the prefix "Ko"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Ko" or variations of it

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

 Creed

-0.73

ingham

-0.71

glass

-0.67

 Attribution

-0.65

 iPads

-0.65

 additive

-0.64

senal

-0.64

ibly

-0.63

IBLE

-0.62

 fingertips

-0.62

POSITIVE LOGITS

osta

1.07

ji

1.03

zzi

1.02

eln

1.02

pps

1.01

Å¡

0.99

unin

0.99

jo

0.94

etter

0.94

lyak

0.94

Activations Density 0.019%

names related to the prefix "Ko"

mentions of the name "Ko" or variations of it

No Comments

No Known Activations