INDEX

Explanations

references to the amount or level of something, especially emphasizing high amounts

oai_token-act-pair · gpt-3.5-turbo

references to quantities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

boa

-0.67

ARP

-0.65

stal

-0.63

INA

-0.60

lez

-0.60

 Roots

-0.60

stood

-0.60

 Pill

-0.59

 paras

-0.59

nee

-0.58

POSITIVE LOGITS

 quantity

1.37

 quantities

1.17

itatively

1.00

Quantity

0.98

quant

0.92

ĸļ

0.86

ickr

0.83

ategory

0.79

odynam

0.76

iful

0.75

Activations Density 0.005%

references to the amount or level of something, especially emphasizing high amounts

references to quantities

No Comments

No Known Activations