INDEX

Explanations

references to specific individuals in news articles

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

names of prominent media figures and political personalities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

 Sakura

-0.63

 convent

-0.57

 Agric

-0.57

 Dominion

-0.56

 completion

-0.55

 Trooper

-0.55

mine

-0.54

 Hearth

-0.53

Nek

-0.53

PSP

-0.52

POSITIVE LOGITS

Jr

0.91

(@

0.88

oÄŁ

0.84

andowski

0.82

enegger

0.78

aka

0.71

 joked

0.71

frey

0.70

steen

0.70

agher

0.69

Activations Density 0.194%

references to specific individuals in news articles

names of prominent media figures and political personalities

No Comments

No Known Activations