INDEX

Explanations

phrases related to breaking news

oai_token-act-pair · gpt-3.5-turbo

terms associated with breaking news and updates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 nostalg

-0.75

vag

-0.70

adm

-0.65

 expressive

-0.62

inav

-0.61

ufact

-0.61

 challeng

-0.60

SHIP

-0.59

 monop

-0.59

joining

-0.59

POSITIVE LOGITS

 Emails

1.14

urrency

0.75

 Latest

0.67

 Coverage

0.67

 Desk

0.67

 Brief

0.67

Desk

0.65

 Debug

0.65

NEWS

0.64

Mub

0.64

Activations Density 0.055%

phrases related to breaking news

terms associated with breaking news and updates

No Comments

No Known Activations