INDEX

Explanations

phrases indicating specific individuals or groups within a given context

oai_token-act-pair · gpt-3.5-turbo

phrases involving the word "who" and references to specific capacities or roles within contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

axy

-0.69

illion

-0.67

TPS

-0.66

skinned

-0.66

´

-0.65

å¸

-0.65

person

-0.64

 pione

-0.63

mage

-0.63

mouth

-0.63

POSITIVE LOGITS

 Canaver

0.84

 Register

0.73

 Regulations

0.72

 Bord

0.66

 resp

0.63

ndra

0.63

 Governments

0.61

 Cert

0.61

lishes

0.61

 roundup

0.61

Activations Density 0.314%

phrases indicating specific individuals or groups within a given context

phrases involving the word "who" and references to specific capacities or roles within contexts

No Comments

No Known Activations