INDEX

Explanations

references to joint actions, collaborations, or partnerships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

æ¾

-0.17

ÑģÑĤÑĢÐ¾Ð²

-0.15

ÏĦÎ¹ÎºÎ¿

-0.15

 Ø³Ø§Ø²ÛĮ

-0.14

ophage

-0.14

 Ø¯Ø§ÙĪ

-0.14

ØŃØ¯Ø§Ø«

-0.14

atos

-0.14

empre

-0.14

atorium

-0.14

POSITIVE LOGITS

 venture

0.36

venture

0.29

ure

0.29

 Venture

0.27

 ventures

0.26

-stock

0.23

 effort

0.22

 forces

0.22

 custody

0.22

ures

0.21

Activations Density 0.016%

references to joint actions, collaborations, or partnerships

No Comments

No Known Activations