INDEX

Explanations

references to entities or concepts represented by the letter 'Y'

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter 'Y'

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

 Prelude

-0.68

 heads

-0.67

 Wonderland

-0.67

 arresting

-0.62

 Creed

-0.61

 coupled

-0.60

 Prol

-0.58

Sno

-0.57

 prevail

-0.57

 Inquis

-0.56

POSITIVE LOGITS

von

1.25

vette

1.19

onge

1.17

ield

1.13

ves

1.13

ousse

1.12

usra

1.06

arb

1.03

orks

1.02

ulia

1.01

Activations Density 0.022%

references to entities or concepts represented by the letter 'Y'

occurrences of the letter 'Y'

No Comments

No Known Activations