INDEX

Explanations

references to social interactions and gatherings in crowded settings

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

Incre

-0.18

OKIE

-0.14

Ymd

-0.14

orado

-0.14

ASSERT

-0.14

orman

-0.14

olle

-0.13

aN

-0.13

ellar

-0.13

Gatt

-0.13

POSITIVE LOGITS

 people

0.27

people

0.23

 souls

0.20

 bodies

0.20

 guests

0.20

äººçļĦ

0.19

 peoples

0.19

 visitors

0.19

 folks

0.18

äººä»¬

0.18

Activations Density 0.013%

references to social interactions and gatherings in crowded settings

No Comments

No Known Activations