INDEX

Explanations

conditional phrases and examples related to instructions or scenarios

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

bjerg

-0.16

ifo

-0.15

 táºŃp

-0.14

WEEN

-0.14

 mouth

-0.14

arine

-0.13

Ð°Ð¼Ð¿

-0.13

 brush

-0.13

 upside

-0.13

POSITIVE LOGITS

 ÙħØ«ÙĦØ§

0.17

 hypoth

0.15

PCODE

0.15

rÃ¡m

0.14

Ð¾Ð³Ð°

0.14

åŃĺæ¡£

0.14

(dl

0.13

 putas

0.13

=center

0.13

.dw

0.13

Activations Density 0.045%

conditional phrases and examples related to instructions or scenarios

No Comments

No Known Activations