INDEX

Explanations

phrases related to governmental and military subjects

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

ba

-0.15

 Moff

-0.14

Ba

-0.14

 Gover

-0.14

Bd

-0.14

opt

-0.14

MF

-0.14

_cached

-0.14

tron

-0.14

 Spect

-0.14

POSITIVE LOGITS

åĲ¹

0.18

iere

0.17

iert

0.17

IDGE

0.16

ople

0.16

ibling

0.16

lom

0.16

 Brothers

0.15

bre

0.15

arser

0.15

Activations Density 0.023%

phrases related to governmental and military subjects

No Comments

No Known Activations