INDEX

Explanations

specific numerical patterns and measurements

oai_token-act-pair · gpt-3.5-turbo

numerical data and statistics related to time, particularly years and percentages

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

ument

-0.87

ensibly

-0.76

acre

-0.70

esse

-0.69

esc

-0.69

enium

-0.69

aminer

-0.68

uberty

-0.66

rison

-0.64

ropri

-0.63

POSITIVE LOGITS

 Oral

0.68

Aer

0.67

 Pharaoh

0.66

Amount

0.66

aints

0.65

 Kart

0.62

terms

0.62

 Azerb

0.62

 Kuala

0.61

 flights

0.61

Activations Density 0.062%

specific numerical patterns and measurements

numerical data and statistics related to time, particularly years and percentages

No Comments

No Known Activations