INDEX

Explanations

phrases related to sitting or being in a particular position or scenario

oai_token-act-pair · gpt-3.5-turbo

phrases related to the act of sitting or being positioned in relation to something or someone

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

catentry

-0.79

Ĥİ

-0.76

qua

-0.70

esi

-0.66

ella

-0.64

ģ«

-0.64

 endeavour

-0.64

Ħ¢

-0.63

eli

-0.62

£ı

-0.61

POSITIVE LOGITS

 idle

0.77

isode

0.70

 seiz

0.68

ocene

0.68

oin

0.67

 pedest

0.66

ulhu

0.65

rolled

0.63

 desks

0.63

 unused

0.62

Activations Density 0.125%

phrases related to sitting or being in a particular position or scenario

phrases related to the act of sitting or being positioned in relation to something or someone

No Comments

No Known Activations