INDEX

Explanations

characters or symbols associated with specific languages

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

jÃŃ

-0.16

loi

-0.15

sole

-0.15

wayne

-0.15

erosis

-0.15

olist

-0.14

*(*

-0.14

clus

-0.14

pii

-0.14

..↵↵↵↵

-0.13

POSITIVE LOGITS

ľ

0.18

 uninitialized

0.17

¸

0.17

ļ

0.17

±

0.16

Ģ

0.15

¯

0.15

¬

0.15

Ĺ

0.15

ħ

0.15

Activations Density 0.004%

characters or symbols associated with specific languages

No Comments

No Known Activations