INDEX

Explanations

proper nouns, such as names of people, places, and organizations

oai_token-act-pair · gpt-3.5-turbo

specific names, terms, and phrases relevant to unique topics or concepts

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

 respectively

-0.72

anwhile

-0.63

 Tradable

-0.59

 bilateral

-0.57

Interstitial

-0.55

kW

-0.55

 guiActiveUnfocused

-0.54

çĶ°

-0.54

IOC

-0.53

 Ambro

-0.52

POSITIVE LOGITS

 anymore

0.78

agar

0.71

coin

0.61

hack

0.59

 spoilers

0.59

ratom

0.57

escape

0.56

tan

0.56

berries

0.54

:(

0.53

Activations Density 1.411%

proper nouns, such as names of people, places, and organizations

specific names, terms, and phrases relevant to unique topics or concepts

No Comments

No Known Activations