INDEX

Explanations

concepts related to understanding and explaining ideas

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

ivec

-0.17

ivre

-0.16

ture

-0.16

uros

-0.15

hiba

-0.14

ontent

-0.14

íĮ¨

-0.14

illac

-0.14

 Ð¾Ð³Ð»Ñı

-0.13

 Woodward

-0.13

POSITIVE LOGITS

as

0.35

how

0.34

 regarding

0.27

how

0.25

 about

0.25

 what

0.21

 vá»ģ

0.21

 cÃ³mo

0.21

 ÐºÐ°Ðº

0.21

 regard

0.20

Activations Density 0.236%

concepts related to understanding and explaining ideas

No Comments

No Known Activations