INDEX

Explanations

repeated instances of a particular character or symbol

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

kaar

-0.18

obel

-0.16

olas

-0.16

obus

-0.16

hek

-0.15

oldt

-0.15

isos

-0.15

beam

-0.15

olv

-0.15

amas

-0.14

POSITIVE LOGITS

wi

0.22

wie

0.22

wiÄĻ

0.21

led

0.20

rod

0.20

ÅĽ

0.20

lad

0.18

ro

0.18

WI

0.18

wiÄħ

0.17

Activations Density 0.003%

repeated instances of a particular character or symbol

No Comments

No Known Activations