INDEX

Explanations

terms related to grading systems and records management

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

iffe

-0.15

 Yunan

-0.15

hta

-0.15

anguard

-0.14

aria

-0.14

.volley

-0.14

-0.13

Lif

-0.13

raf

-0.13

POSITIVE LOGITS

 contest

0.15

 contests

0.14

olls

0.14

gee

0.14

/Runtime

0.14

irut

0.13

.Atomic

0.13

eed

0.13

azon

0.13

amilia

0.13

Activations Density 0.125%

terms related to grading systems and records management

No Comments

No Known Activations