INDEX

Explanations

phrases related to providing guidance or instructions

oai_token-act-pair · gpt-3.5-turbo

references to guides or instructional materials

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 spect

-0.73

yss

-0.68

 Anniversary

-0.66

 anniversary

-0.65

activated

-0.64

ellen

-0.62

mercial

-0.62

pite

-0.61

ulz

-0.61

outh

-0.61

POSITIVE LOGITS

 guide

1.32

guide

1.28

 guides

1.24

 Guides

1.02

 Guide

1.02

book

0.99

books

0.98

Guide

0.90

posts

0.88

tips

0.87

Activations Density 0.010%

phrases related to providing guidance or instructions

references to guides or instructional materials

No Comments

No Known Activations