INDEX

Explanations

syntax constructs, particularly those that indicate code structure or control flow, such as function calls and conditionals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.18.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

 +=↵

-0.14

 Loose

-0.13

 fitted

-0.13

asma

-0.13

walker

-0.13

Î»Î¹Î¬

-0.13

Rew

-0.13

ance

-0.13

 defensive

-0.13

POSITIVE LOGITS

Cao

0.15

áº£o

0.15

echa

0.14

razione

0.14

adj

0.14

 scal

0.14

eden

0.14

orde

0.14

kÃ¶

0.13

jeta

0.13

Activations Density 0.105%

syntax constructs, particularly those that indicate code structure or control flow, such as function calls and conditionals

No Comments

No Known Activations