INDEX

Explanations

phrases indicating comparative emphasis, particularly those using "as" and "much."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

 closer

-0.16

onz

-0.15

ounder

-0.15

 ganz

-0.15

 licz

-0.15

ÏģÎµ

-0.15

ayer

-0.14

ÑĢÐ¾Ñĩ

-0.14

igger

-0.14

eper

-0.14

POSITIVE LOGITS

 much

0.77

much

0.71

 Much

0.66

Much

0.62

 MUCH

0.60

 mucho

0.52

 beaucoup

0.34

 viel

0.32

 veel

0.31

 mycket

0.28

Activations Density 0.072%

phrases indicating comparative emphasis, particularly those using "as" and "much."

No Comments

No Known Activations