INDEX

Explanations

names of locations, specifically "Santa Clara" and "Alameda"

oai_token-act-pair · gpt-3.5-turbo

the name "Clara," specifically 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

unct

-0.76

ramid

-0.72

mediated

-0.72

ileaks

-0.71

lopp

-0.70

Marginal

-0.67

nuclear

-0.64

Western

-0.64

PUT

-0.64

draft

-0.64

POSITIVE LOGITS

 Clara

1.49

 Oswald

0.92

asma

0.81

 Barton

0.80

iffe

0.79

 valley

0.77

phia

0.75

Lens

0.73

 Jenna

0.72

 County

0.72

Activations Density 0.006%

names of locations, specifically "Santa Clara" and "Alameda"

the name "Clara," specifically in various contexts

No Comments

No Known Activations