INDEX

Explanations

questions about specific terms or concepts

oai_token-act-pair · gpt-3.5-turbo

questions or definitions regarding concepts and terminology

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 regrets

-0.85

 regretted

-0.73

 regret

-0.72

 fearful

-0.69

bably

-0.69

ersen

-0.67

 awaited

-0.66

 panicked

-0.65

 wiser

-0.65

POSITIVE LOGITS

 Basically

1.36

 Essentially

1.35

Definition

1.30

Basically

1.29

 Definition

1.25

 Definitions

1.19

 Basics

1.09

 nutshell

1.09

Overview

1.07

Wikipedia

0.97

Activations Density 0.399%

questions about specific terms or concepts

questions or definitions regarding concepts and terminology

No Comments

No Known Activations