INDEX

Explanations

references to law enforcement and police-related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

 Guards

-0.19

adin

-0.19

 guards

-0.15

elman

-0.15

Å¡t

-0.15

dot

-0.15

optgroup

-0.15

à¹Ģà¸¥à¸¢

-0.14

iras

-0.14

 policemen

-0.14

POSITIVE LOGITS

 canine

0.22

 Narc

0.22

 Gang

0.21

 Collision

0.20

 Burg

0.20

 Evidence

0.19

Ã¥n

0.19

Sex

0.18

 narc

0.18

 narcotics

0.18

Activations Density 0.060%

references to law enforcement and police-related terms

No Comments

No Known Activations