INDEX

Explanations

names of specific geographical locations

oai_token-act-pair · gpt-3.5-turbo

references to specific geographical locations or cities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

emin

-0.71

cou

-0.64

 Advertisement

-0.64

taboola

-0.64

Bow

-0.63

 Lethal

-0.62

 Corpor

-0.62

Cou

-0.60

 Officer

-0.60

LAR

-0.60

POSITIVE LOGITS

stadt

4.94

esville

1.23

ingen

1.12

etsk

1.04

 Petersburg

1.04

inese

1.02

analysis

1.00

arde

0.98

aston

0.93

zhou

0.91

Activations Density 0.036%

names of specific geographical locations

references to specific geographical locations or cities

No Comments

No Known Activations