INDEX

Explanations

emotionally charged expressions or sentiments

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

 ÑģÑĤÐ¾ÑĢÐ¾Ð½

-0.15

Ø§ÙĪÙĬ

-0.15

vang

-0.14

seau

-0.14

inne

-0.14

 gameTime

-0.14

ãĤĤãģĨ

-0.14

VERR

-0.14

ascus

-0.14

 ÑģÑĤÐ¾ÑĢ

-0.13

POSITIVE LOGITS

 regard

0.23

 regards

0.19

 respect

0.19

 permission

0.17

 Ð¿Ð¾Ð¼Ð¾ÑīÑĮÑİ

0.17

 whom

0.17

intent

0.17

respect

0.15

uth

0.15

outh

0.15

Activations Density 0.013%

emotionally charged expressions or sentiments

No Comments

No Known Activations