INDEX

Explanations

phrases that include various forms of "have."

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

:CGRect

-0.15

ible

-0.15

cio

-0.15

sville

-0.15

imd

-0.14

mind

-0.14

appa

-0.14

Æ°á»¡ng

-0.14

.pointer

-0.14

contres

-0.14

POSITIVE LOGITS

 options

0.21

 choices

0.20

 until

0.19

 OPTIONS

0.17

Until

0.17

Choices

0.17

 hwnd

0.17

 responsibilities

0.17

 Options

0.16

 permission

0.16

Activations Density 0.107%

phrases that include various forms of "have."

No Comments

No Known Activations