INDEX

Explanations

descriptions of routes or pathways

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

 upstream

-0.16

Forward

-0.15

ulk

-0.15

 FORWARD

-0.15

 Forward

-0.14

alex

-0.14

à¸¢à¸§

-0.14

 onward

-0.14

 Heading

-0.13

 nearer

-0.13

POSITIVE LOGITS

 almost

0.19

Interrupt

0.18

 nearly

0.18

 roughly

0.17

interrupt

0.17

Parallel

0.16

almost

0.16

AAP

0.15

 unint

0.15

 alongside

0.15

Activations Density 0.052%

descriptions of routes or pathways

No Comments

No Known Activations