INDEX

Explanations

actions or phrases related to user interactions and convenience, particularly involving clicks and voice commands

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

iser

-0.16

angep

-0.14

/is

-0.14

 genital

-0.14

pitch

-0.14

_CY

-0.14

undi

-0.13

Ð°ÑĢÑĩ

-0.13

à¸ªà¸Ķ

-0.13

fri

-0.13

POSITIVE LOGITS

 merely

0.17

 mere

0.16

cery

0.15

inox

0.15

akh

0.15

mere

0.15

_Native

0.14

Ð²Ð¾Ð´

0.14

toJson

0.14

iano

0.14

Activations Density 0.031%

actions or phrases related to user interactions and convenience, particularly involving clicks and voice commands

No Comments

No Known Activations