INDEX

Explanations

numerical values and names in various contexts, particularly in titles and ratings

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

 Hobby

-0.16

ä¸²

-0.16

ifest

-0.15

æłĳ

-0.15

IFF

-0.14

 fandom

-0.14

ogh

-0.14

opoulos

-0.14

áº¥p

-0.14

 Canter

-0.14

POSITIVE LOGITS

Tin

0.17

 Virtual

0.16

Virtual

0.15

Tou

0.15

 virtual

0.15

Publisher

0.14

ascar

0.14

 sublic

0.14

apur

0.14

æĭŁ

0.14

Activations Density 0.089%

numerical values and names in various contexts, particularly in titles and ratings

No Comments

No Known Activations