INDEX

Explanations

references to abstract concepts and symbols

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

íģ¼

-0.15

æļ

-0.14

 weer

-0.13

@g

-0.13

hry

-0.13

eron

-0.12

emm

-0.12

 Fiat

-0.12

encion

-0.12

 uncomment

-0.12

POSITIVE LOGITS

oplay

0.17

 ÎŁÏħ

0.16

 Nisan

0.16

 Ãĸl

0.15

 Optical

0.15

 OnInit

0.15

Nonce

0.15

.pageX

0.14

 optical

0.14

PointerType

0.14

Activations Density 0.080%

references to abstract concepts and symbols

No Comments

No Known Activations