INDEX

Explanations

phrases related to news alerts

oai_token-act-pair · gpt-3.5-turbo

keywords related to alerts and notifications, particularly in a news context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

construct

-0.81

lad

-0.78

ocative

-0.77

apy

-0.72

assemb

-0.70

gro

-0.70

ple

-0.69

ardless

-0.68

arist

-0.68

acting

-0.67

POSITIVE LOGITS

 Alert

1.02

 alerts

0.92

 forecast

0.73

 subscriptions

0.72

chini

0.72

 alert

0.70

 derail

0.69

 subscription

0.68

iflower

0.66

 Subscribe

0.64

Activations Density 0.018%

phrases related to news alerts

keywords related to alerts and notifications, particularly in a news context

No Comments

No Known Activations