INDEX

Explanations

proper names and specific historical references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 12-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L12R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.12.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

SizeMode

-0.16

ãĥĨãĥ«

-0.15

rott

-0.15

ÏīÎ´

-0.15

:variables

-0.14

andel

-0.14

illiseconds

-0.14

$MESS

-0.14

Ø´Ø¯

-0.14

rellas

-0.14

POSITIVE LOGITS

ÑĥÑĢÐ½

0.17

 Mang

0.15

ounters

0.15

 Invent

0.14

 seating

0.14

 partially

0.14

 Philippe

0.13

pt

0.13

 fraction

0.13

Activations Density 0.169%

proper names and specific historical references

No Comments

No Known Activations