INDEX

Explanations

affirmative references to group inclusivity

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vez

-0.19

æģ©

-0.16

abis

-0.15

unden

-0.15

 è©³ç´°

-0.15

 stuff

-0.14

Ø±Ø§

-0.14

ylie

-0.14

reen

-0.14

ucch

-0.13

POSITIVE LOGITS

 manner

0.21

 available

0.18

 Available

0.17

 sorts

0.17

tel

0.17

 hell

0.16

 AVAILABLE

0.15

available

0.15

 four

0.15

Activations Density 0.071%

affirmative references to group inclusivity

No Comments

No Known Activations