INDEX

Explanations

references to utility services and their related contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

 usual

-0.18

alim

-0.18

uzzi

-0.18

ures

-0.17

_utils

-0.16

ald

-0.16

-utils

-0.16

_util

-0.15

gor

-0.15

.au

-0.15

POSITIVE LOGITS

isateur

0.27

itarian

0.27

fully

0.22

ful

0.22

ties

0.19

FUL

0.19

fulness

0.18

belt

0.18

full

0.17

 belt

0.17

Activations Density 0.023%

references to utility services and their related contexts

No Comments

No Known Activations