INDEX

Explanations

ways to achieve solutions or improvements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 16-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L16R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.16.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

lez

-0.14

legg

-0.14

ethyl

-0.14

ulings

-0.14

dued

-0.13

/problem

-0.13

mÃ¼r

-0.13

 fare

-0.13

(extra

-0.13

.Hex

-0.13

POSITIVE LOGITS

 ways

0.34

 solution

0.30

way

0.29

 solutions

0.27

solution

0.24

 methods

0.24

 Ways

0.23

 alternate

0.22

 method

0.22

 alternative

0.22

Activations Density 0.241%

ways to achieve solutions or improvements

No Comments

No Known Activations