INDEX

Explanations

instances of shared or universal approaches and concepts

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

CAB

-0.15

ours

-0.15

olo

-0.14

quest

-0.14

Ð¾Ð»Ð¾ÑĤ

-0.14

åłĨ

-0.14

á»ħ

-0.14

breadcrumb

-0.14

ÐµÑģÑĮ

-0.13

anders

-0.13

POSITIVE LOGITS

Across

0.18

phin

0.18

APTER

0.17

0.16

 Across

0.15

ë¡±

0.15

0.14

 across

0.14

onya

0.14

 reused

0.14

Activations Density 0.310%

instances of shared or universal approaches and concepts

No Comments

No Known Activations

instances of shared or universal approaches and concepts

No Comments

No Known Activations