INDEX

Explanations

discussions about programming solutions and workarounds

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

ubs

-0.16

 Dillon

-0.16

atsby

-0.16

ullo

-0.15

bole

-0.14

Injected

-0.14

ieme

-0.14

ardin

-0.14

ilot

-0.14

_chk

-0.13

POSITIVE LOGITS

 instead

0.19

 resort

0.19

 simply

0.18

 workaround

0.17

 create

0.17

instead

0.17

 SOLUTION

0.15

kl

0.15

use

0.15

 Instead

0.15

Activations Density 0.185%

discussions about programming solutions and workarounds

No Comments

No Known Activations