INDEX

Explanations

instances of the word "and" and phrases that indicate collaboration or togetherness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

ccount

-0.15

eld

-0.15

ippi

-0.14

adelphia

-0.14

éļĨ

-0.14

ByPrimaryKey

-0.13

Ã©li

-0.13

eny

-0.13

ClickListener

-0.13

hur

-0.13

POSITIVE LOGITS

lay

0.54

Lay

0.47

lay

0.44

 laying

0.44

 laid

0.40

 lays

0.38

Lie

0.37

 lying

0.35

lie

0.34

LAY

0.32

Activations Density 0.036%

instances of the word "and" and phrases that indicate collaboration or togetherness

No Comments

No Known Activations