INDEX

Explanations

mentions of locations, particularly cities like San Francisco and San Diego

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ï¸ı

-0.72

 numbering

-0.70

 numbered

-0.68

llers

-0.68

ilater

-0.67

llor

-0.65

 monop

-0.64

ICS

-0.63

pus

-0.62

hower

-0.62

POSITIVE LOGITS

 Francisco

1.26

 Diego

1.22

ctuary

1.13

 Antonio

1.11

San

1.09

 Bernardino

1.02

ibel

0.97

itary

0.97

Disk

0.97

gha

0.93

Activations Density 0.018%

mentions of locations, particularly cities like San Francisco and San Diego

No Comments

No Known Activations