INDEX

Explanations

locations, specifically capital cities

oai_token-act-pair · gpt-3.5-turbo

mentions of geographic capitals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

potion

-0.90

AUT

-0.73

chrom

-0.71

 Reincarn

-0.71

wcsstore

-0.65

vent

-0.64

piracy

-0.62

nir

-0.62

âķĲâķĲ

-0.62

External

-0.62

POSITIVE LOGITS

 city

1.05

 Sana

1.00

 metropolitan

0.97

 Beirut

0.95

 Tripoli

0.94

 Baghdad

0.93

 Ankara

0.92

 Damascus

0.92

 Islamabad

0.91

 Kabul

0.91

Activations Density 0.029%

locations, specifically capital cities

mentions of geographic capitals

No Comments

No Known Activations