INDEX

Explanations

conjunctions, specifically instances of "and" and "or."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 2-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L2R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.2.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

pz

-0.15

utow

-0.15

ãĥ³ãĤ¬

-0.15

.RowCount

-0.15

robat

-0.15

 erotische

-0.15

pawn

-0.14

çĤī

-0.14

 annonces

-0.14

/***/

-0.13

POSITIVE LOGITS

/or

0.21

ibel

0.14

icken

0.14

0.13

ala

0.13

 Prom

0.13

endon

0.13

 Picker

0.13

yr

0.13

 Grove

0.13

Activations Density 1.156%

conjunctions, specifically instances of "and" and "or."

No Comments

No Known Activations