INDEX

Explanations

significant numerical or quantitative information related to items or elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

CBS

-0.16

CBS

-0.15

 Jackson

-0.15

 armored

-0.14

/Gate

-0.14

 rumor

-0.14

 gate

-0.14

 theaters

-0.14

 favors

-0.14

licken

-0.13

POSITIVE LOGITS

Os

0.25

os

0.23

 Drawable

0.22

osu

0.22

/os

0.21

Beat

0.21

Bindable

0.20

 Beat

0.20

Colour

0.20

Drawable

0.19

Activations Density 0.001%

significant numerical or quantitative information related to items or elements

No Comments

No Known Activations