INDEX

Explanations

dates or time-related phrases

oai_token-act-pair · gpt-3.5-turbo

keywords and phrases related to advertisements and updates

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

 destro

-0.85

 eleph

-0.76

 neighb

-0.76

xual

-0.72

nesday

-0.70

undai

-0.69

hement

-0.69

 estranged

-0.68

 overth

-0.67

 occas

-0.67

POSITIVE LOGITS

Contribut

1.04

 largeDownload

1.03

Posted

1.02

Trivia

0.97

Details

0.94

Contents

0.93

Interested

0.93

Product

0.93

rawdownloadcloneembedreportprint

0.92

ARM

0.91

Activations Density 0.071%

dates or time-related phrases

keywords and phrases related to advertisements and updates

No Comments

No Known Activations