INDEX

Explanations

references to television and TV-related content

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

resher

-0.18

/Peak

-0.16

alg

-0.15

uggage

-0.15

æĹıèĩªæ²»

-0.15

ivist

-0.15

inho

-0.15

Äĥ

-0.15

ivism

-0.15

icher

-0.15

POSITIVE LOGITS

ÙĬÙĪÙĨ

0.19

/movie

0.18

onda

0.17

orca

0.17

iland

0.16

Ã¡z

0.15

/video

0.15

olution

0.15

/print

0.15

-show

0.14

Activations Density 0.024%

references to television and TV-related content

No Comments

No Known Activations