INDEX

Explanations

mentions of the term "President-elect" with high activation values

oai_token-act-pair · gpt-3.5-turbo

references to a political figure and their electoral status

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

MpServer

-0.67

ENE

-0.64

 glim

-0.63

Definition

-0.60

cms

-0.60

 ingred

-0.60

 endemic

-0.60

 diamonds

-0.60

 informants

-0.58

$$

-0.58

POSITIVE LOGITS

 Donald

1.34

 Trump

1.21

Donald

1.11

 TRUMP

1.06

realDonaldTrump

1.04

elect

1.04

trump

1.03

Trump

1.00

 Rodrigo

0.99

 Barack

0.96

Activations Density 0.040%

mentions of the term "President-elect" with high activation values

references to a political figure and their electoral status

No Comments

No Known Activations