INDEX

Explanations

phrases that emphasize significance or primary importance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

.digest

-0.16

 possibility

-0.15

íĬ

-0.14

 handful

-0.13

iams

-0.13

ients

-0.13

duit

-0.13

alo

-0.13

lier

-0.13

elpers

-0.13

POSITIVE LOGITS

 benefits

0.22

 ways

0.21

 questions

0.20

 beaut

0.20

 often

0.19

 problems

0.19

 challenges

0.19

 beauty

0.19

 constant

0.18

ardin

0.18

Activations Density 0.063%

phrases that emphasize significance or primary importance

No Comments

No Known Activations