INDEX

Explanations

references to new developments and changes in systems, organizations, or policies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

herits

-0.17

 renew

-0.15

URA

-0.14

chÃ¡z

-0.14

uren

-0.14

existent

-0.14

adÄ±ÄŁÄ±

-0.14

orsche

-0.14

etik

-0.13

Ã¼lÃ¼k

-0.13

POSITIVE LOGITS

iche

0.18

icho

0.15

-found

0.15

 _↵

0.14

igers

0.14

Ð¤ÐĽ

0.14

iger

0.13

 bÃ¬

0.13

æİª

0.13

arr

0.13

Activations Density 0.055%

references to new developments and changes in systems, organizations, or policies

No Comments

No Known Activations