INDEX

Explanations

punctuations and numbers that are out of context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

uling

-0.16

weets

-0.15

QUOTE

-0.15

roje

-0.14

ulers

-0.14

bling

-0.14

remium

-0.14

ayload

-0.13

 Grund

-0.13

POSITIVE LOGITS

ilden

0.19

åĭ

0.19

oba

0.16

ksen

0.16

 Sabb

0.16

RTL

0.15

yles

0.15

auer

0.14

ÑĤÑĢÐ¾

0.14

ne

0.14

Activations Density 0.002%

punctuations and numbers that are out of context

No Comments

No Known Activations