INDEX

Explanations

elements related to writing and composition techniques

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 12-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L12R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.12.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

zos

-0.19

uyen

-0.16

setFlash

-0.15

laps

-0.14

HAL

-0.14

 poloÅ¾

-0.14

idges

-0.14

ÑĥÑģÑĤÐ°

-0.14

ingu

-0.14

zbo

-0.14

POSITIVE LOGITS

pon

0.16

jo

0.16

etta

0.15

pur

0.15

Wi

0.14

aires

0.14

pun

0.13

ussen

0.13

pa

0.13

 prox

0.13

Activations Density 0.024%

elements related to writing and composition techniques

No Comments

No Known Activations