INDEX

Explanations

concepts related to selection and choice-making

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

ë¨¸ëĭĪ

-0.16

 ÑģÐ¾Ð±Ð¾Ð¹

-0.15

ayan

-0.14

Å¡li

-0.14

mlink

-0.14

igue

-0.13

iore

-0.13

ÏİÎ½

-0.13

 tiener

-0.13

emiz

-0.13

POSITIVE LOGITS

 suitable

0.27

 appropriate

0.25

 candidates

0.24

 candidate

0.22

 Suitable

0.20

uitable

0.20

appropriate

0.20

/select

0.20

 sites

0.19

 partners

0.19

Activations Density 0.135%

concepts related to selection and choice-making

No Comments

No Known Activations