INDEX

Explanations

instances of attempts or actions related to criminal behavior and dishonesty

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 12-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L12R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.12.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

juana

-0.15

 longevity

-0.14

ResultsController

-0.14

icon

-0.14

ãĥĪãĥ«

-0.14

 dÃ¼r

-0.14

/banner

-0.13

 kategor

-0.13

UMP

-0.13

asil

-0.13

POSITIVE LOGITS

inet

0.17

isay

0.17

opus

0.17

onet

0.16

igne

0.15

 Harm

0.15

activex

0.15

GetX

0.15

ideon

0.14

 Plans

0.14

Activations Density 0.315%

instances of attempts or actions related to criminal behavior and dishonesty

No Comments

No Known Activations