INDEX

Explanations

the word "all" in various contexts, particularly when it appears in phrases of emphasis or completeness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 27-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L27R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.27.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

roz

-0.21

Î»Ïī

-0.17

ãĥĥãĥģ

-0.17

acie

-0.16

arness

-0.16

vang

-0.16

æľ¬

-0.15

edom

-0.15

isle

-0.15

-BEGIN

-0.15

POSITIVE LOGITS

 sorts

0.41

 kinds

0.39

 manner

0.32

 KIND

0.26

of

0.26

 sort

0.25

 SORT

0.23

sort

0.23

 types

0.21

 those

0.21

Activations Density 0.069%

the word "all" in various contexts, particularly when it appears in phrases of emphasis or completeness

No Comments

No Known Activations