INDEX

Explanations

references to specific names and titles related to cultural works

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

urga

-0.19

urse

-0.16

iets

-0.15

 mbox

-0.14

ousse

-0.14

het

-0.14

okes

-0.14

urses

-0.13

habi

-0.13

 Ð²Ðº

-0.13

POSITIVE LOGITS

arella

0.16

Î±ÏģÎ±

0.14

PRO

0.14

åĸ

0.14

amba

0.14

 brom

0.14

[Index

0.14

ynos

0.14

iang

0.14

_INTERFACE

0.14

Activations Density 0.011%

references to specific names and titles related to cultural works

No Comments

No Known Activations