INDEX

Explanations

references to violence and attacks in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 23-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L23R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.23.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

 cannons

-0.15

oce

-0.15

EventType

-0.15

rane

-0.15

åħĭæĸ¯

-0.14

scr

-0.13

zw

-0.13

Łèĥ½

-0.13

CRET

-0.13

POSITIVE LOGITS

 innocent

0.42

 targets

0.38

 Innoc

0.33

 Targets

0.33

 innoc

0.32

 helpless

0.30

targets

0.30

 civilians

0.29

 innocence

0.28

 target

0.28

Activations Density 0.301%

references to violence and attacks in various contexts

No Comments

No Known Activations