INDEX

Explanations

words related to placeholders or pending information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 27-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L27R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.27.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

æľĭ

-0.20

alth

-0.19

çĶ

-0.16

à¹Ĥà¸¢

-0.15

ç±³

-0.15

aje

-0.15

ãĥĥãĥī

-0.15

izzo

-0.15

 å±

-0.14

generation

-0.14

POSITIVE LOGITS

-cross

0.21

 cross

0.21

cross

0.19

 Cross

0.19

 CROSS

0.18

Cross

0.18

_cross

0.16

 crossed

0.16

 Crossing

0.15

 sher

0.14

Activations Density 0.028%

words related to placeholders or pending information

No Comments

No Known Activations