INDEX

Explanations

phrases indicating positions of influence or control within various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

ntl

-0.18

bero

-0.16

stoupil

-0.16

æĹ¦

-0.15

Prostit

-0.15

ziel

-0.14

pn

-0.14

etto

-0.14

stoup

-0.14

iali

-0.14

POSITIVE LOGITS

kdo

0.17

 Governors

0.15

 TMPro

0.15

ÑĢÐµÑģÑģ

0.14

ordes

0.14

igator

0.14

.lv

0.13

igid

0.13

Î½Î¿Ïį

0.13

refixer

0.13

Activations Density 0.128%

phrases indicating positions of influence or control within various contexts

No Comments

No Known Activations