INDEX

Explanations

instances where an article or piece of text is complete

oai_token-act-pair · gpt-3.5-turbo

strong indicators of urgency or significant events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 tradem

-0.79

 amnesty

-0.75

 reper

-0.74

thodox

-0.74

undai

-0.74

 exting

-0.73

 eleph

-0.70

ilty

-0.70

 acquies

-0.69

 destro

-0.69

POSITIVE LOGITS

CLOSE

1.15

WASHINGTON

1.07

Trivia

1.06

Abstract

1.05

Still

1.05

Description

1.04

Overview

1.03

Story

1.02

Untitled

1.01

Breaking

1.01

Activations Density 0.116%

instances where an article or piece of text is complete

strong indicators of urgency or significant events

No Comments

No Known Activations