INDEX

Explanations

references to names and titles, particularly in the context of creative works

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

iyon

-0.16

 ÑģÑĩ

-0.15

ervo

-0.15

Ã¦kke

-0.15

#ac

-0.15

essor

-0.15

obb

-0.15

SSIP

-0.15

ì²¨ë¶Ģ

-0.14

implify

-0.14

POSITIVE LOGITS

ist

0.16

pon

0.15

Kod

0.15

pos

0.15

 align

0.15

Activations Density 0.004%

references to names and titles, particularly in the context of creative works

No Comments

No Known Activations