INDEX

Explanations

mentions of the word "Merc" or variations such as "Merciful" or "Merciless"

oai_token-act-pair · gpt-3.5-turbo

references to the term "Merc" and its variations, possibly related to a character or entity named Merc in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

doors

-0.86

eking

-0.79

FORMATION

-0.77

feat

-0.69

growth

-0.68

åĤ

-0.68

Traditional

-0.68

cki

-0.67

chy

-0.66

Lei

-0.66

POSITIVE LOGITS

enaries

1.24

iless

1.16

 simultane

1.13

 Merc

1.02

oun

0.97

uria

0.94

 mosqu

0.94

 occas

0.92

ifully

0.92

 merc

0.91

Activations Density 0.003%

mentions of the word "Merc" or variations such as "Merciful" or "Merciless"

references to the term "Merc" and its variations, possibly related to a character or entity named Merc in various contexts

No Comments

No Known Activations