INDEX

Explanations

punctuation marks, particularly periods and questions

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

asley

-0.17

.chapter

-0.15

Dak

-0.15

isiyle

-0.14

shaw

-0.14

sworth

-0.14

 Sche

-0.14

 Story

-0.14

ammed

-0.13

POSITIVE LOGITS

iyan

0.16

edom

0.15

ardi

0.14

chter

0.14

Ð¿Ð°ÑĤ

0.14

 Containers

0.13

 Dorm

0.13

zeros

0.13

hte

0.13

Ã¤ft

0.13

Activations Density 0.026%

punctuation marks, particularly periods and questions

No Comments

No Known Activations