INDEX

Explanations

keywords related to urgent or extreme situations and actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 5-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L5R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.5.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

ecer

-0.15

WithPath

-0.15

Ñľ

-0.14

elim

-0.14

ivil

-0.14

ellan

-0.14

CTest

-0.14

riba

-0.14

Truthy

-0.14

iyon

-0.14

POSITIVE LOGITS

 Talk

0.17

Talk

0.15

war

0.14

rieg

0.14

enne

0.14

à¤¾à¤¦à¤¨

0.13

 coordin

0.13

 Survival

0.13

 Gian

0.13

 resolve

0.13

Activations Density 0.032%

keywords related to urgent or extreme situations and actions

No Comments

No Known Activations