INDEX

Explanations

references to prison and criminal activity

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

arin

-0.16

à¸¸à¹Ī

-0.15

á»Ń

-0.15

 Foreign

-0.14

anners

-0.14

lico

-0.14

aar

-0.14

 foreign

-0.14

Foreign

-0.14

ÏħÎ½Î±

-0.14

POSITIVE LOGITS

 prison

0.54

 inmate

0.48

 prisoner

0.47

 Prison

0.46

 prisoners

0.45

 inmates

0.45

 prisons

0.41

 jail

0.37

 incarcerated

0.33

 Corrections

0.33

Activations Density 0.173%

references to prison and criminal activity

No Comments

No Known Activations