INDEX

Explanations

notifications or communications that are formal or legally significant

oai_token-act-pair · gpt-3.5-turbo

instances of the word "notice."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ECD

-0.70

 rall

-0.69

UES

-0.67

 corrid

-0.67

 cheek

-0.65

UFF

-0.64

âĹ¼

-0.63

EVA

-0.63

 ancest

-0.61

 percussion

-0.61

POSITIVE LOGITS

andum

0.96

 notice

0.94

 notices

0.88

anty

0.82

 Notice

0.82

itarian

0.81

isance

0.81

 Signs

0.81

aban

0.78

pee

0.78

Activations Density 0.012%

notifications or communications that are formal or legally significant

instances of the word "notice."

No Comments

No Known Activations