INDEX

Explanations

phrases related to directing attention or effort towards specific tasks or goals

oai_token-act-pair · gpt-3.5-turbo

phrases focusing on policy recommendations and priorities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

DragonMagazine

-0.86

©¶æ¥µ

-0.83

listed

-0.79

soever

-0.78

UP

-0.76

soType

-0.75

nown

-0.73

zynski

-0.70

opter

-0.70

ns

-0.70

POSITIVE LOGITS

 improving

1.36

 maximizing

1.22

 optimizing

1.14

 preserving

1.13

 achieving

1.08

 solving

1.07

 repairing

1.07

 enhancing

1.06

 minimizing

1.05

 ensuring

1.04

Activations Density 0.122%

phrases related to directing attention or effort towards specific tasks or goals

phrases focusing on policy recommendations and priorities

No Comments

No Known Activations