INDEX

Explanations

proper nouns, particularly names or brands

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

eling

-0.19

elli

-0.18

idon

-0.17

elly

-0.17

elle

-0.17

elson

-0.16

es

-0.15

ese

-0.15

ellan

-0.15

et

-0.15

POSITIVE LOGITS

entine

0.28

entina

0.26

uable

0.25

uation

0.24

leys

0.23

entin

0.23

uations

0.23

uetype

0.23

val

0.22

=val

0.21

Activations Density 0.025%

proper nouns, particularly names or brands

No Comments

No Known Activations