INDEX

Explanations

phrases that emphasize connections between different themes 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 @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

à¹Ģà¸īà¸ŀà¸²à¸°

-0.15

TransparentColor

-0.15

 klim

-0.14

 altri

-0.14

ading

-0.14

alytics

-0.13

.Criteria

-0.13

ilet

-0.13

ä¼´

-0.13

iane

-0.13

POSITIVE LOGITS

 oppos

0.29

 worlds

0.29

 disparate

0.28

 sexes

0.27

 halves

0.27

two

0.26

 opposing

0.24

 incompatible

0.24

 disciplines

0.24

 different

0.23

Activations Density 0.285%

phrases that emphasize connections between different themes or elements

No Comments

No Known Activations