INDEX

Explanations

location names and geographic features

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

 Governors

-0.17

AtPath

-0.16

otine

-0.16

å¾Ħ

-0.15

tube

-0.15

eros

-0.15

ãİ¡

-0.14

ÄĽÅ¾

-0.14

 governors

-0.14

 Isles

-0.13

POSITIVE LOGITS

 division

0.26

 district

0.25

 city

0.22

division

0.21

 Division

0.20

 divisions

0.20

 districts

0.19

ensi

0.18

 town

0.18

 collector

0.18

Activations Density 0.063%

location names and geographic features

No Comments

No Known Activations