INDEX

Explanations

mentions of the Punjabi culture or language

oai_token-act-pair · gpt-3.5-turbo

individual letters or characters that appear frequently in context

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 pent

-0.59

ashtra

-0.56

 brace

-0.56

 proceeds

-0.56

cape

-0.55

_-

-0.55

 dividends

-0.54

akespeare

-0.53

 termin

-0.53

govtrack

-0.53

POSITIVE LOGITS

hement

0.80

 Antar

0.78

 Nieto

0.69

insky

0.69

Cola

0.68

aneers

0.64

ishers

0.64

lic

0.62

entials

0.61

ansky

0.61

Activations Density 0.163%

mentions of the Punjabi culture or language

individual letters or characters that appear frequently in context

No Comments

No Known Activations