INDEX

Explanations

names of various news outlets or newspapers

oai_token-act-pair · gpt-3.5-turbo

indicators of events and actions in news or media contexts

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

apon

-0.79

 hiber

-0.72

 ageing

-0.72

 capitals

-0.69

hid

-0.67

 eleven

-0.66

 moaning

-0.64

 alone

-0.63

apons

-0.63

 tram

-0.63

POSITIVE LOGITS

Buy

1.47

window

1.21

CLOSE

1.03

FILE

0.93

hello

0.87

 FILE

0.83

Study

0.80

PHOTOS

0.79

quire

0.79

Aug

0.78

Activations Density 0.031%

names of various news outlets or newspapers

indicators of events and actions in news or media contexts

No Comments

No Known Activations