INDEX

Explanations

keywords related to news articles or content recommendation

oai_token-act-pair · gpt-3.5-turbo

references to related content or topics

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

»Ĵ

-0.78

naissance

-0.72

Ĥİ

-0.72

anut

-0.71

ainted

-0.71

esan

-0.68

ãĥ´ãĤ¡

-0.67

UGE

-0.67

elfth

-0.66

ucket

-0.66

POSITIVE LOGITS

 Stories

1.23

 Articles

1.16

 Links

1.11

 Content

1.09

 Topics

1.00

 articles

0.98

 Posts

0.96

 Story

0.96

 ARTICLE

0.96

 Video

0.95

Activations Density 0.030%

keywords related to news articles or content recommendation

references to related content or topics

No Comments

No Known Activations