INDEX

Explanations

words related to switching or toggling states or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 13-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L13R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.13.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

strict

-0.16

eus

-0.15

uly

-0.15

icious

-0.15

à¸£

-0.15

exion

-0.14

untas

-0.14

sson

-0.14

ifice

-0.14

POSITIVE LOGITS

ero

0.20

esa

0.19

et

0.18

aroo

0.17

Endian

0.16

sit

0.16

etu

0.16

INCREMENT

0.15

pery

0.15

tower

0.15

Activations Density 0.053%

words related to switching or toggling states or actions

No Comments

No Known Activations