INDEX

Explanations

references to prisons and correctional systems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 27-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L27R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.27.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

à¸¸à¹Ī

-0.15

ëł¹

-0.14

_sampler

-0.14

aar

-0.14

_Device

-0.13

utr

-0.13

Ð°ÑĤÐ¸

-0.13

Ã©ric

-0.13

 sublic

-0.13

æ»ħ

-0.13

POSITIVE LOGITS

 prison

0.76

 Prison

0.68

 prisoner

0.65

 prisoners

0.63

 jail

0.61

 inmate

0.61

 prisons

0.59

 inmates

0.57

 Jail

0.53

 Correction

0.50

Activations Density 0.317%

references to prisons and correctional systems

No Comments

No Known Activations