INDEX

Explanations

mentions of favorite things or items

oai_token-act-pair · gpt-3.5-turbo

references to favorite things, particularly in the context of films, characters, and other personal preferences

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

iless

-0.78

udic

-0.74

amination

-0.74

psey

-0.70

ividual

-0.70

ĸļ

-0.69

inarily

-0.68

arching

-0.66

idan

-0.65

 Canaver

-0.65

POSITIVE LOGITS

 Oscar

0.82

RPG

0.77

ORPG

0.74

kid

0.73

ãĤ¼

0.72

ãĤ´ãĥ³

0.71

 Grammy

0.70

aceae

0.69

 Occupations

0.67

Pop

0.67

Activations Density 0.252%

mentions of favorite things or items

references to favorite things, particularly in the context of films, characters, and other personal preferences

No Comments

No Known Activations