INDEX

Explanations

numerical data and references to competitors or specific people

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

()->

-0.15

 walnut

-0.14

 Ð±Ð¾Ðº

-0.14

zig

-0.14

æª

-0.13

å°ıå§Ĳ

-0.13

izen

-0.13

Box

-0.13

å´

-0.13

 SCRIPT

-0.13

POSITIVE LOGITS

 sprint

0.25

 cross

0.24

 Sprint

0.22

 nord

0.20

 relay

0.19

Ski

0.19

 Nordic

0.19

 distance

0.19

ski

0.19

cross

0.18

Activations Density 0.010%

numerical data and references to competitors or specific people

No Comments

No Known Activations