INDEX

Explanations

references to acts of crime or wrongdoing

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

 Superior

-0.15

 attacker

-0.15

 offending

-0.15

thern

-0.14

illa

-0.14

 attackers

-0.13

ilo

-0.13

_compress

-0.13

 possibility

-0.13

/ay

-0.13

POSITIVE LOGITS

 acts

0.23

acts

0.22

 suicide

0.21

 treason

0.19

 genocide

0.19

ocide

0.18

 sudoku

0.17

 adultery

0.17

 burglary

0.17

Ð°Ð´Ñĥ

0.17

Activations Density 0.021%

references to acts of crime or wrongdoing

No Comments

No Known Activations