INDEX

Explanations

numerical dates and comment counts related to posts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

pur

-0.17

ittel

-0.15

ph

-0.15

phem

-0.15

ANGO

-0.14

Ð´Ð°Ñħ

-0.14

lator

-0.14

bm

-0.14

çİ°

-0.14

inee

-0.14

POSITIVE LOGITS

 geil

0.19

 Categories

0.16

 Auth

0.16

 Author

0.15

chwitz

0.15

ervo

0.15

olet

0.15

heits

0.14

 geile

0.14

orrect

0.14

Activations Density 0.005%

numerical dates and comment counts related to posts

No Comments

No Known Activations