INDEX

Explanations

structured lists or itemized points in a document

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

ãģıãĤī

-0.15

leftright

-0.14

abay

-0.14

ÃŃc

-0.14

-FIRST

-0.14

celik

-0.13

usi

-0.13

edReader

-0.13

Ø§ØŃÛĮ

-0.13

 [âĢ¦]↵↵

-0.13

POSITIVE LOGITS

Å¡ek

0.16

entai

0.16

 ÐĳÐ¾Ð»

0.15

Ð¼Ð¸ÑĢ

0.15

atik

0.15

 Î£Ïĩ

0.14

aran

0.14

elan

0.14

 Lambert

0.13

 cheer

0.13

Activations Density 0.027%

structured lists or itemized points in a document

No Comments

No Known Activations