INDEX

Explanations

references to scientific or legal principles and terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

XCT

-0.14

 handleMessage

-0.14

à¸±à¸Ķ

-0.14

ìĿ´íĦ°

-0.14

å½ķ

-0.13

alnÄ±z

-0.13

imin

-0.13

alian

-0.13

Ø¨ÙĪØ§Ø³Ø·Ø©

-0.13

 BindingFlags

-0.13

POSITIVE LOGITS

 says

0.52

 states

0.49

say

0.43

 saying

0.41

 Says

0.40

says

0.39

states

0.37

 à¤ķà¤¹à¤¤

0.36

 stating

0.35

 said

0.34

Activations Density 0.390%

references to scientific or legal principles and terminology

No Comments

No Known Activations