INDEX

Explanations

relational words indicating comparison or contradiction

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

èĥ½å¤Ł

-0.21

æĽ´åĬł

-0.15

åħ·æľī

-0.14

 maniÃ¨re

-0.14

jes

-0.13

hatt

-0.13

onth

-0.13

ãģ§ãģĤãģ£ãģŁ

-0.13

ä¹ĭåĲİ

-0.13

stoff

-0.13

POSITIVE LOGITS

 combos

0.17

ãģ¾ãģļ

0.15

 EITHER

0.14

piler

0.14

gov

0.14

ilon

0.13

Ã£

0.13

ÃŃs

0.13

oloji

0.13

yes

0.13

Activations Density 0.116%

relational words indicating comparison or contradiction

No Comments

No Known Activations