INDEX

Explanations

terms associated with sexual misconduct and legal terminology around it

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

 latter

-0.16

ylko

-0.16

ancies

-0.16

phans

-0.15

ients

-0.15

iae

-0.15

REFIX

-0.15

webtoken

-0.15

ÅĦ

-0.14

gars

-0.14

POSITIVE LOGITS

odore

0.33

adays

0.29

atre

0.28

atomy

0.24

xiety

0.22

quarters

0.21

gether

0.21

etheless

0.20

ÑįÑĤÐ¾Ð¼Ñĥ

0.20

tlement

0.19

Activations Density 1.468%

terms associated with sexual misconduct and legal terminology around it

No Comments

No Known Activations