INDEX

Explanations

words related to rankings or numerical positions

oai_token-act-pair · gpt-3.5-turbo

references to ordinal numbers, particularly those related to seasons or rankings

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

Statistics

-0.71

ICO

-0.67

NUM

-0.66

++++++++++++++++

-0.65

EFF

-0.64

Money

-0.64

bda

-0.63

bers

-0.62

 Keane

-0.62

uploads

-0.61

POSITIVE LOGITS

 inning

1.18

ority

0.97

ieth

0.95

irty

0.88

umbnail

0.87

awk

0.86

eenth

0.82

teenth

0.82

ulhu

0.81

orne

0.80

Activations Density 0.025%

words related to rankings or numerical positions

references to ordinal numbers, particularly those related to seasons or rankings

No Comments

No Known Activations