INDEX

Explanations

references to political positions and titles, particularly those involving "prime."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

ëŀ

-0.16

neau

-0.16

weg

-0.15

(es

-0.15

aines

-0.14

enate

-0.14

ÃŃsticas

-0.13

anken

-0.13

 Cort

-0.13

ieu

-0.13

POSITIVE LOGITS

 mover

0.15

ayer

0.15

ps

0.15

è»

0.15

0.14

Ø§Ø¯Ø§

0.14

ordial

0.14

erea

0.14

ayers

0.13

voices

0.13

Activations Density 0.032%

references to political positions and titles, particularly those involving "prime."

No Comments

No Known Activations