INDEX

Explanations

references to social divisions and human relationships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

ulumi

-0.15

.rl

-0.15

 å¾

-0.15

idla

-0.14

 wanting

-0.14

 needing

-0.14

edition

-0.14

.bs

-0.13

figcaption

-0.13

reated

-0.13

POSITIVE LOGITS

who

0.33

wh

0.28

qui

0.26

 that

0.26

who

0.24

 yang

0.24

 which

0.23

Who

0.23

que

0.22

 ÐºÐ¾ÑĤÐ¾ÑĢÑĭÐ¹

0.22

Activations Density 0.238%

references to social divisions and human relationships

No Comments

No Known Activations