INDEX

Explanations

dates in a specific format

oai_token-act-pair · gpt-3.5-turbo

structured lists or enumerations in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ancies

-0.69

ral

-0.69

conn

-0.67

bara

-0.65

"},"

-0.65

rams

-0.64

bos

-0.63

rations

-0.62

ammed

-0.62

ades

-0.61

POSITIVE LOGITS

 Firstly

0.88

 introdu

0.71

 Introduction

0.69

 Sabha

0.68

 Understand

0.67

 Brief

0.66

 Overview

0.66

 Name

0.65

 Introdu

0.65

apo

0.65

Activations Density 0.365%

dates in a specific format

structured lists or enumerations in the text

No Comments

No Known Activations