INDEX

Explanations

references to technology or specific technical terms

oai_token-act-pair · gpt-3.5-turbo

phrases related to local news or events

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

 amnesty

-0.72

thodox

-0.70

ilty

-0.67

 intervened

-0.67

 acquies

-0.66

 reper

-0.65

utic

-0.65

 entitle

-0.63

lling

-0.63

 embargo

-0.63

POSITIVE LOGITS

Trivia

0.97

CLOSE

0.97

WASHINGTON

0.96

Still

0.94

Quote

0.91

Posted

0.91

Media

0.90

About

0.89

Overview

0.88

CTV

0.88

Activations Density 0.095%

references to technology or specific technical terms

phrases related to local news or events

No Comments

No Known Activations