INDEX

Explanations

instances of quantified or specific numerical information and conditions

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

rez

-0.19

RuntimeObject

-0.18

udeau

-0.17

rz

-0.15

pla

-0.15

WER

-0.15

ouve

-0.15

é¡µéĿ¢åŃĺæ¡£å¤ĩä»½

-0.15

ovah

-0.15

ekli

-0.14

POSITIVE LOGITS

 ones

0.28

 Moss

0.17

 Ones

0.17

Arb

0.16

sir

0.14

ness

0.14

Sir

0.14

roj

0.14

 Goldberg

0.14

ebek

0.14

Activations Density 0.226%

instances of quantified or specific numerical information and conditions

No Comments

No Known Activations