INDEX

Explanations

commands or suggestions related to giving voice or agency to oneself or others

oai_token-act-pair · gpt-3.5-turbo

expressions of agency and empowerment regarding individual choices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

enegger

-1.11

Downloadha

-0.76

ModLoader

-0.73

 Annotations

-0.72

ALD

-0.66

interstitial

-0.65

HAM

-0.64

lia

-0.64

imilar

-0.62

 Medal

-0.62

POSITIVE LOGITS

 breathe

0.78

 shine

0.76

YP

0.75

unic

0.74

 expire

0.69

 suffice

0.68

 aside

0.67

orce

0.66

ghai

0.66

 loose

0.65

Activations Density 0.081%

commands or suggestions related to giving voice or agency to oneself or others

expressions of agency and empowerment regarding individual choices

No Comments

No Known Activations