INDEX

Explanations

technical terms and identifiers, especially in computer programming and music contexts

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

-esque

-0.18

keit

-0.16

 láº½

-0.15

_DOM

-0.15

eum

-0.15

ÑģÑĮ

-0.15

sh

-0.14

nger

-0.14

èĬĤ

-0.14

Ú¯Ùĩ

-0.14

POSITIVE LOGITS

emiah

0.19

stakes

0.18

headed

0.18

emony

0.17

Ð¾Ð¿Ð°Ñģ

0.16

DED

0.15

Ã©ny

0.15

ĶåĽŀ

0.15

iatrics

0.15

lify

0.15

Activations Density 3.751%

technical terms and identifiers, especially in computer programming and music contexts

No Comments

No Known Activations