INDEX

Explanations

mentions of broadcasters or broadcasting organizations

oai_token-act-pair · gpt-3.5-turbo

references to news and media organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

othes

-0.75

illard

-0.74

udes

-0.70

urat

-0.68

reenshots

-0.68

arcity

-0.66

cession

-0.65

hots

-0.64

plets

-0.62

 Venice

-0.62

POSITIVE LOGITS

 broadcaster

1.42

 broadcasters

1.15

 announcer

0.91

caster

0.86

 broadcasting

0.85

 licensee

0.85

 Broadcast

0.84

 Broadcasting

0.84

 Seym

0.83

levision

0.82

Activations Density 0.010%

mentions of broadcasters or broadcasting organizations

references to news and media organizations

No Comments

No Known Activations