INDEX

Explanations

mentions of UCLA

oai_token-act-pair · gpt-3.5-turbo

references to UCLA and Oscar-related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

odic

-0.79

BOOK

-0.75

anooga

-0.73

piece

-0.67

bits

-0.66

DonaldTrump

-0.66

ipper

-0.66

ĩ

-0.65

ively

-0.64

ote

-0.64

POSITIVE LOGITS

 Bruins

0.87

 Wilde

0.77

enthal

0.76

 Kobe

0.76

 ANGEL

0.75

 UCLA

0.74

ITED

0.74

 Pictures

0.74

 Schwarzenegger

0.70

 Pist

0.69

Activations Density 0.076%

mentions of UCLA

references to UCLA and Oscar-related terms

No Comments

No Known Activations