INDEX

Explanations

references to unknown identities and undisclosed information

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

MediaType

-0.16

iggs

-0.16

nder

-0.15

orton

-0.14

istrict

-0.14

oloj

-0.13

deo

-0.13

donnees

-0.13

ekyll

-0.13

meer

-0.13

POSITIVE LOGITS

nor

0.24

(éĩĳ

0.18

Nor

0.17

nor

0.17

abi

0.16

Nor

0.16

wdx

0.15

(æľĪ

0.15

ensi

0.15

.visitMethod

0.14

Activations Density 0.098%

references to unknown identities and undisclosed information

No Comments

No Known Activations