INDEX

Explanations

phrases that repeatedly reference specific groups or individuals as "those."

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

enstein

-0.17

ayne

-0.15

ä¸»äºº

-0.14

aptops

-0.14

those

-0.14

 outset

-0.14

Ë

-0.14

à¥ģà¤°à¤¸

-0.14

idan

-0.13

isode

-0.13

POSITIVE LOGITS

who

0.42

who

0.32

 whom

0.29

 same

0.29

curity

0.28

Who

0.26

Who

0.24

 kinds

0.23

 Ø§ÙĦØ°ÙĬÙĨ

0.23

 whose

0.22

Activations Density 0.060%

phrases that repeatedly reference specific groups or individuals as "those."

No Comments

No Known Activations