INDEX

Explanations

various forms of the word "wrap" and associated actions or conditions related to physical objects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

fold

-0.16

arti

-0.15

otten

-0.14

arges

-0.14

zik

-0.14

antis

-0.13

quier

-0.13

 Dion

-0.13

agher

-0.13

xia

-0.13

POSITIVE LOGITS

plemented

0.17

-from

0.16

æ¸Ī

0.15

ervers

0.15

aupt

0.14

iegel

0.14

ogram

0.14

pling

0.14

 ná»ģn

0.13

ted

0.13

Activations Density 0.176%

various forms of the word "wrap" and associated actions or conditions related to physical objects

No Comments

No Known Activations