INDEX

Explanations

symbols related to questions and punctuation marks

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

bsite

-0.15

orman

-0.14

 Ø§ÙĦØ¨ÙĦØ¯

-0.14

ÅĻez

-0.14

ascal

-0.14

AndGet

-0.13

Ø¨ÙĦ

-0.13

iveness

-0.13

bum

-0.13

agle

-0.13

POSITIVE LOGITS

ipar

0.18

Answer

0.17

-answer

0.16

 erotische

0.15

 Answer

0.15

 pione

0.15

ainen

0.15

—↵↵

0.14

ulares

0.14

trieve

0.14

Activations Density 0.089%

symbols related to questions and punctuation marks

No Comments

No Known Activations