INDEX

Explanations

words or phrases related to specific identities or nationalities

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

illet

-0.17

 evil

-0.17

olk

-0.16

ilded

-0.15

dana

-0.15

icious

-0.15

 Evil

-0.15

omic

-0.15

auer

-0.15

evil

-0.14

POSITIVE LOGITS

regular

0.24

replace

0.22

Ir

0.21

irr

0.21

ving

0.21

ir

0.21

ritable

0.20

relevant

0.19

ises

0.19

Ã©n

0.18

Activations Density 0.019%

words or phrases related to specific identities or nationalities

No Comments

No Known Activations