INDEX

Explanations

words related to size or scale

oai_token-act-pair · gpt-3.5-turbo

terms related to size

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

VPN

-0.77

DCS

-0.73

ance

-0.69

roma

-0.69

INA

-0.69

Fre

-0.68

ËĪ

-0.67

Nou

-0.67

Blade

-0.66

ario

-0.65

POSITIVE LOGITS

 sized

1.14

 eleph

0.91

 proport

0.90

 millenn

0.88

sized

0.83

 scale

0.82

 proportions

0.79

owship

0.76

 conflic

0.76

 practition

0.76

Activations Density 0.009%

words related to size or scale

terms related to size

No Comments

No Known Activations