INDEX

Explanations

specific locations or place names

oai_token-act-pair · gpt-3.5-turbo

instances of the term "LOS" and its variations, indicating a focus on the Los Angeles area or related discussions

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

Ab

-0.77

ating

-0.67

ated

-0.65

ate

-0.64

ations

-0.63

 vanilla

-0.63

acci

-0.63

 Shard

-0.63

aba

-0.62

essert

-0.60

POSITIVE LOGITS

LOS

4.63

los

1.81

Los

1.27

LO

1.19

LES

1.13

SAN

1.12

ESA

1.11

CLA

1.03

IAS

1.01

LA

1.01

Activations Density 0.034%

specific locations or place names

instances of the term "LOS" and its variations, indicating a focus on the Los Angeles area or related discussions

No Comments

No Known Activations