INDEX

Explanations

elements related to programming or coding

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

Not in Any Lists

No Comments

Negative Logits

ede

-0.17

 males

-0.17

 male

-0.17

anson

-0.16

éĽĦ

-0.15

à¸Ļà¸²à¸¢

-0.15

ew

-0.15

dro

-0.15

male

-0.15

 Profession

-0.14

POSITIVE LOGITS

ufen

0.18

 addCriterion

0.17

iversit

0.16

 civilian

0.15

 civilians

0.15

sonian

0.15

lemn

0.15

frauen

0.15

Illuminate

0.14

_UNICODE

0.14

Activations Density 1.058%

elements related to programming or coding

No Comments

No Known Activations