INDEX

Explanations

locations, specifically focusing on Los Angeles

oai_token-act-pair · gpt-3.5-turbo

references to Los Angeles

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

tics

-0.74

witch

-0.72

colour

-0.72

aye

-0.70

ilk

-0.69

 upstream

-0.68

ECB

-0.67

rx

-0.67

BOOK

-0.66

 pouch

-0.66

POSITIVE LOGITS

Los

3.44

Los

2.94

LA

2.13

 Pasadena

2.03

 LAPD

1.90

Las

1.76

 Anaheim

1.70

 Sacramento

1.68

San

1.66

 Oakland

1.62

Activations Density 0.017%

locations, specifically focusing on Los Angeles

references to Los Angeles

No Comments

No Known Activations