INDEX

Explanations

programming-related keywords, particularly those associated with function definitions and parameters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

bum

-0.15

chip

-0.15

urt

-0.15

.opens

-0.14

azi

-0.14

à¸©

-0.14

esser

-0.14

obi

-0.14

_:*

-0.14

afen

-0.14

POSITIVE LOGITS

ìĿ´íĦ°

0.16

orida

0.15

/OR

0.15

ousand

0.15

lectual

0.14

swer

0.14

etheless

0.14

ividual

0.13

ascus

0.13

&action

0.13

Activations Density 0.015%

programming-related keywords, particularly those associated with function definitions and parameters

No Comments

No Known Activations