INDEX

Explanations

references to religious teachings or figures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

çĨ

-0.16

okies

-0.16

Ð°Ð»Ð¸Ð·Ð¸

-0.16

eroon

-0.15

pak

-0.15

 å®®

-0.15

akte

-0.15

à¸Ħà¸§

-0.14

ebek

-0.14

esis

-0.14

POSITIVE LOGITS

 narr

0.19

Narrated

0.18

Ans

0.17

 Companion

0.17

 Battles

0.17

 THROW

0.17

uar

0.16

 narration

0.16

Ban

0.16

 tribe

0.16

Activations Density 0.045%

references to religious teachings or figures

No Comments

No Known Activations