INDEX

Explanations

discussions about organizational focus on outcomes and effectiveness rather than just quantity or cost

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 27-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L27R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.27.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

 both

-0.17

NotAllowed

-0.17

 Sanity

-0.16

ugas

-0.14

 genuinely

-0.14

apus

-0.14

 real

-0.14

flation

-0.13

 BOTH

-0.13

real

-0.13

POSITIVE LOGITS

 merely

0.24

 mere

0.24

isol

0.24

mere

0.23

 superficial

0.22

 solely

0.21

 isolated

0.21

 cookie

0.19

 quantity

0.19

åıªæĺ¯

0.18

Activations Density 0.935%

discussions about organizational focus on outcomes and effectiveness rather than just quantity or cost

No Comments

No Known Activations