INDEX

Explanations

references to numerical data and bibliographic information in academic citations

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

lette

-0.15

erb

-0.14

ILLISE

-0.14

 repmat

-0.14

qi

-0.14

compass

-0.14

coder

-0.13

inka

-0.13

-kind

-0.13

atar

-0.13

POSITIVE LOGITS

_LEG

0.16

ied

0.15

orde

0.15

istrovstvÃŃ

0.14

isode

0.13

Ìģc

0.13

angered

0.13

ROUP

0.13

ENE

0.12

uggy

0.12

Activations Density 0.029%

references to numerical data and bibliographic information in academic citations

No Known Activations