INDEX

Explanations

the word "adia" with high activation levels

oai_token-act-pair · gpt-3.5-turbo

terms related to a specific geographic or cultural location, particularly variations of "adia" and "adian."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Noel

-0.75

 Notice

-0.71

 Klopp

-0.69

Jar

-0.69

 utilitarian

-0.68

 sorted

-0.64

 revolver

-0.64

 Checks

-0.63

 Snape

-0.63

 Conserv

-0.63

POSITIVE LOGITS

adia

4.15

adian

3.56

adal

2.10

adium

1.62

iba

1.33

adan

1.25

adesh

1.25

alian

1.24

ades

1.24

enary

1.12

Activations Density 0.024%

the word "adia" with high activation levels

terms related to a specific geographic or cultural location, particularly variations of "adia" and "adian."

No Comments

No Known Activations