INDEX

Explanations

programming-related keywords and structures in code

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

tam

-0.16

esel

-0.15

 Heller

-0.14

alent

-0.14

ÙĦÙĥ

-0.14

olit

-0.14

ivas

-0.14

lings

-0.13

CHANT

-0.13

ole

-0.13

POSITIVE LOGITS

byt

0.14

ache

0.14

uron

0.14

brick

0.14

 Highlands

0.14

etsk

0.14

wert

0.13

edback

0.13

atten

0.13

stash

0.13

Activations Density 0.001%

programming-related keywords and structures in code

No Comments

No Known Activations