INDEX

Explanations

programming-related keywords and structures in code snippets

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

IFrame

Link

Negative Logits

åĿĬ

-0.15

finity

-0.15

 varargin

-0.14

Ay

-0.14

 klin

-0.14

ay

-0.14

ought

-0.14

Ay

-0.13

UGHT

-0.13

POSITIVE LOGITS

 person

0.19

 fruit

0.18

Person

0.17

 employee

0.17

 Person

0.17

 Fruit

0.17

 customer

0.17

 weather

0.16

 fruits

0.16

 hello

0.16

Activations Density 0.418%

programming-related keywords and structures in code snippets

No Known Activations