INDEX

Explanations

phrases or statements indicating consistency or compatibility

oai_token-act-pair · gpt-3.5-turbo

phrases indicating consistency or compatibility

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

book

-0.69

DragonMagazine

-0.64

bands

-0.61

backs

-0.61

bom

-0.60

along

-0.60

 credited

-0.59

cloud

-0.59

chem

-0.58

 hairc

-0.58

POSITIVE LOGITS

 CLIENT

0.72

 achieving

0.67

¥µ

0.67

cellence

0.66

emale

0.66

ensus

0.66

clus

0.63

onential

0.62

ensis

0.62

termination

0.62

Activations Density 0.145%

phrases or statements indicating consistency or compatibility

phrases indicating consistency or compatibility

No Comments

No Known Activations