INDEX

Explanations

the word "Kl" and its variations, often associated with records or classifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

igure

-0.18

illard

-0.17

-0.15

¿

-0.14

arto

-0.14

 Hawth

-0.14

Yer

-0.14

lage

-0.14

ELS

-0.14

urret

-0.13

POSITIVE LOGITS

áº·ng

0.19

itoris

0.17

axon

0.17

Kl

0.16

uger

0.16

IMENT

0.16

 divergence

0.15

atch

0.15

uge

0.15

udget

0.15

Activations Density 0.006%

the word "Kl" and its variations, often associated with records or classifications

No Comments

No Known Activations