INDEX

Explanations

uncommon or surprising experiences

oai_token-act-pair · gpt-3.5-turbo

phrases expressing novelty or uniqueness in experiences

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

alore

-0.69

nels

-0.67

Newsletter

-0.64

 Skydragon

-0.64

pta

-0.63

redients

-0.59

nel

-0.59

Accessory

-0.58

anwhile

-0.58

Spread

-0.57

POSITIVE LOGITS

 anything

1.28

 anybody

1.08

 anyone

1.07

anything

1.03

nor

0.91

 such

0.87

any

0.87

ANY

0.86

 Anything

0.83

 anywhere

0.81

Activations Density 0.156%

uncommon or surprising experiences

phrases expressing novelty or uniqueness in experiences

No Comments

No Known Activations