INDEX

Explanations

specific suffixes or prefixes that denote unique attributes or categories related to various subjects, including quests, events, and political affiliations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

asca

-0.18

Ø¯Ø§ÙĨÙĦÙĪØ¯

-0.15

ROLLER

-0.14

ilm

-0.14

ÐµÑİ

-0.14

ãĥ«ãĥķ

-0.13

oro

-0.13

 Trot

-0.13

iyi

-0.13

enor

-0.13

POSITIVE LOGITS

0.15

angled

0.15

porte

0.15

/-

0.14

 blank

0.14

pee

0.14

anford

0.14

aan

0.14

ucch

0.14

shield

0.13

Activations Density 0.138%

specific suffixes or prefixes that denote unique attributes or categories related to various subjects, including quests, events, and political affiliations

No Comments

No Known Activations