INDEX

Explanations

negations, particularly in the context of recommendations and opinions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

pons

-0.15

 Rage

-0.15

caff

-0.15

 Trident

-0.14

belt

-0.14

iyel

-0.14

utta

-0.14

ivia

-0.14

ait

-0.14

gateway

-0.14

POSITIVE LOGITS

ä»»ä½ķ

0.23

any

0.23

Any

0.19

_any

0.19

Any

0.19

 anymore

0.17

 anything

0.17

oÅĽci

0.17

eza

0.16

 qualquer

0.16

Activations Density 0.285%

negations, particularly in the context of recommendations and opinions

No Comments

No Known Activations