INDEX

Explanations

mentions of individuals appearing on television or in media

oai_token-act-pair · gpt-3.5-turbo

occurrences of people or entities participating in various events or appearances

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 endeavour

-0.73

 relies

-0.61

 maintained

-0.61

 neurot

-0.60

mentation

-0.60

sterdam

-0.59

"}],"

-0.59

}{

-0.59

 opio

-0.59

 notwithstanding

-0.59

POSITIVE LOGITS

 Billboard

0.70

PLA

0.70

artist

0.67

OPLE

0.66

 Playboy

0.66

 Countdown

0.65

 Cele

0.65

Picture

0.64

 Oval

0.63

TAIN

0.62

Activations Density 0.146%

mentions of individuals appearing on television or in media

occurrences of people or entities participating in various events or appearances

No Comments

No Known Activations