INDEX

Explanations

dates or time-related phrases

oai_token-act-pair · gpt-3.5-turbo

instances of the word "being" and variations in contexts of ongoing actions or states

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

<?

-0.69

 Christy

-0.67

Yad

-0.67

LESS

-0.66

 Magicka

-0.66

LinkedIn

-0.65

Tid

-0.65

ntil

-0.64

soDeliveryDate

-0.64

alde

-0.62

POSITIVE LOGITS

 arrival

0.76

 introdu

0.75

 lapse

0.75

 unsuccessful

0.71

 accident

0.68

pection

0.68

 initial

0.68

 embargo

0.67

 arrive

0.67

 inception

0.66

Activations Density 0.597%

dates or time-related phrases

instances of the word "being" and variations in contexts of ongoing actions or states

No Comments

No Known Activations