INDEX

Explanations

dates and time-related words

oai_token-act-pair · gpt-3.5-turbo

the word "the."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

okes

-0.81

FIELD

-0.77

Desk

-0.71

Tickets

-0.71

ibi

-0.71

estyles

-0.70

arian

-0.70

Header

-0.69

 unless

-0.69

itutes

-0.69

POSITIVE LOGITS

 emergence

1.34

 influx

1.33

 advent

1.29

 arrival

1.26

 proliferation

1.23

 rise

1.19

 fact

1.17

 resurgence

1.16

 increasing

1.11

 absence

1.10

Activations Density 0.261%

dates and time-related words

the word "the."

No Comments

No Known Activations