INDEX

Explanations

the word "even."

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

 even

-0.40

 tháºŃm

-0.34

 EVEN

-0.34

 Even

-0.33

çĶļèĩ³

-0.32

 Ð´Ð°Ð¶Ðµ

-0.31

even

-0.30

Even

-0.30

 nawet

-0.30

 sogar

-0.27

POSITIVE LOGITS

-handed

0.26

 though

0.25

though

0.21

ness

0.21

Though

0.20

 worse

0.20

 Though

0.19

-number

0.19

 remotely

0.18

 partially

0.18

Activations Density 0.046%

the word "even."

No Comments

No Known Activations