INDEX

Explanations

phrases related to a powerful force or entity

oai_token-act-pair · gpt-3.5-turbo

references to juggling or related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Spa

-0.82

 Koen

-0.78

 Ambro

-0.77

Filename

-0.76

earcher

-0.72

 Reyes

-0.70

 Aden

-0.69

 plaque

-0.69

SPONSORED

-0.69

sein

-0.68

POSITIVE LOGITS

 jugg

3.93

 juggling

2.29

 Jugg

1.82

 Mighty

1.41

ighty

1.40

jug

1.07

 mighty

1.04

 grapp

1.03

glers

1.02

 balancing

1.02

Activations Density 0.043%

phrases related to a powerful force or entity

references to juggling or related terms

No Comments

No Known Activations