INDEX

Explanations

phrases that convey contingent or conditional 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

er

-0.16

dea

-0.16

ader

-0.16

ide

-0.16

ses

-0.15

boa

-0.15

sp

-0.14

YLES

-0.14

Ved

-0.14

iras

-0.14

POSITIVE LOGITS

 upon

0.33

upon

0.28

 Upon

0.27

Upon

0.25

ents

0.20

 whether

0.18

ently

0.18

sole

0.17

 whereabouts

0.17

ľĺ

0.17

Activations Density 0.015%

phrases that convey contingent or conditional relationships

No Comments

No Known Activations